orq-ai-sdk 4.2.0rc28__py3-none-any.whl → 4.2.6__py3-none-any.whl

orq_ai_sdk/models/createagentrequestop.py

@@ -90,6 +90,22 @@ class ResponseFormatJSONSchema(BaseModel):
     strict: Optional[bool] = False
     r"""Whether to enable strict schema adherence when generating the output. If set to true, the model will always follow the exact schema defined in the schema field. Only a subset of JSON Schema is supported when strict is true."""
 
+    @model_serializer(mode="wrap")
+    def serialize_model(self, handler):
+        optional_fields = set(["description", "schema", "strict"])
+        serialized = handler(self)
+        m = {}
+
+        for n, f in type(self).model_fields.items():
+            k = f.alias or n
+            val = serialized.get(k)
+
+            if val != UNSET_SENTINEL:
+                if val is not None or k not in optional_fields:
+                    m[k] = val
+
+        return m
+
 
 class JSONSchemaTypedDict(TypedDict):
     r"""
@@ -212,6 +228,22 @@ class StreamOptions(BaseModel):
     include_usage: Optional[bool] = None
     r"""If set, an additional chunk will be streamed before the data: [DONE] message. The usage field on this chunk shows the token usage statistics for the entire request, and the choices field will always be an empty array. All other chunks will also include a usage field, but with a null value."""
 
+    @model_serializer(mode="wrap")
+    def serialize_model(self, handler):
+        optional_fields = set(["include_usage"])
+        serialized = handler(self)
+        m = {}
+
+        for n, f in type(self).model_fields.items():
+            k = f.alias or n
+            val = serialized.get(k)
+
+            if val != UNSET_SENTINEL:
+                if val is not None or k not in optional_fields:
+                    m[k] = val
+
+        return m
+
 
 ThinkingTypedDict = TypeAliasType(
     "ThinkingTypedDict",
@@ -254,6 +286,22 @@ class ToolChoice2(BaseModel):
     type: Optional[ToolChoiceType] = None
     r"""The type of the tool. Currently, only function is supported."""
 
+    @model_serializer(mode="wrap")
+    def serialize_model(self, handler):
+        optional_fields = set(["type"])
+        serialized = handler(self)
+        m = {}
+
+        for n, f in type(self).model_fields.items():
+            k = f.alias or n
+            val = serialized.get(k)
+
+            if val != UNSET_SENTINEL:
+                if val is not None or k not in optional_fields:
+                    m[k] = val
+
+        return m
+
 
 ToolChoice1 = Literal[
     "none",
@@ -312,6 +360,154 @@ class CreateAgentRequestModelConfigurationGuardrails(BaseModel):
     r"""Determines whether the guardrail runs on the input (user message) or output (model response)."""
 
 
+class FallbacksTypedDict(TypedDict):
+    model: str
+    r"""Fallback model identifier"""
+
+
+class Fallbacks(BaseModel):
+    model: str
+    r"""Fallback model identifier"""
+
+
+class RetryTypedDict(TypedDict):
+    r"""Retry configuration for the request"""
+
+    count: NotRequired[float]
+    r"""Number of retry attempts (1-5)"""
+    on_codes: NotRequired[List[float]]
+    r"""HTTP status codes that trigger retry logic"""
+
+
+class Retry(BaseModel):
+    r"""Retry configuration for the request"""
+
+    count: Optional[float] = 3
+    r"""Number of retry attempts (1-5)"""
+
+    on_codes: Optional[List[float]] = None
+    r"""HTTP status codes that trigger retry logic"""
+
+    @model_serializer(mode="wrap")
+    def serialize_model(self, handler):
+        optional_fields = set(["count", "on_codes"])
+        serialized = handler(self)
+        m = {}
+
+        for n, f in type(self).model_fields.items():
+            k = f.alias or n
+            val = serialized.get(k)
+
+            if val != UNSET_SENTINEL:
+                if val is not None or k not in optional_fields:
+                    m[k] = val
+
+        return m
+
+
+ModelConfigurationType = Literal["exact_match",]
+
+
+class CacheTypedDict(TypedDict):
+    r"""Cache configuration for the request."""
+
+    type: ModelConfigurationType
+    ttl: NotRequired[float]
+    r"""Time to live for cached responses in seconds. Maximum 259200 seconds (3 days)."""
+
+
+class Cache(BaseModel):
+    r"""Cache configuration for the request."""
+
+    type: ModelConfigurationType
+
+    ttl: Optional[float] = 1800
+    r"""Time to live for cached responses in seconds. Maximum 259200 seconds (3 days)."""
+
+    @model_serializer(mode="wrap")
+    def serialize_model(self, handler):
+        optional_fields = set(["ttl"])
+        serialized = handler(self)
+        m = {}
+
+        for n, f in type(self).model_fields.items():
+            k = f.alias or n
+            val = serialized.get(k)
+
+            if val != UNSET_SENTINEL:
+                if val is not None or k not in optional_fields:
+                    m[k] = val
+
+        return m
+
+
+LoadBalancerType = Literal["weight_based",]
+
+
+class CreateAgentRequestLoadBalancerModelsTypedDict(TypedDict):
+    model: str
+    r"""Model identifier for load balancing"""
+    weight: NotRequired[float]
+    r"""Weight assigned to this model for load balancing"""
+
+
+class CreateAgentRequestLoadBalancerModels(BaseModel):
+    model: str
+    r"""Model identifier for load balancing"""
+
+    weight: Optional[float] = 0.5
+    r"""Weight assigned to this model for load balancing"""
+
+    @model_serializer(mode="wrap")
+    def serialize_model(self, handler):
+        optional_fields = set(["weight"])
+        serialized = handler(self)
+        m = {}
+
+        for n, f in type(self).model_fields.items():
+            k = f.alias or n
+            val = serialized.get(k)
+
+            if val != UNSET_SENTINEL:
+                if val is not None or k not in optional_fields:
+                    m[k] = val
+
+        return m
+
+
+class LoadBalancer1TypedDict(TypedDict):
+    type: LoadBalancerType
+    models: List[CreateAgentRequestLoadBalancerModelsTypedDict]
+
+
+class LoadBalancer1(BaseModel):
+    type: LoadBalancerType
+
+    models: List[CreateAgentRequestLoadBalancerModels]
+
+
+LoadBalancerTypedDict = LoadBalancer1TypedDict
+r"""Load balancer configuration for the request."""
+
+
+LoadBalancer = LoadBalancer1
+r"""Load balancer configuration for the request."""
+
+
+class TimeoutTypedDict(TypedDict):
+    r"""Timeout configuration to apply to the request. If the request exceeds the timeout, it will be retried or fallback to the next model if configured."""
+
+    call_timeout: float
+    r"""Timeout value in milliseconds"""
+
+
+class Timeout(BaseModel):
+    r"""Timeout configuration to apply to the request. If the request exceeds the timeout, it will be retried or fallback to the next model if configured."""
+
+    call_timeout: float
+    r"""Timeout value in milliseconds"""
+
+
 class ParametersTypedDict(TypedDict):
     r"""Model behavior parameters that control how the model generates responses. Common parameters: `temperature` (0-1, randomness), `max_completion_tokens` (max output length), `top_p` (sampling diversity). Advanced: `frequency_penalty`, `presence_penalty`, `response_format` (JSON/structured), `reasoning_effort`, `seed` (reproducibility). Support varies by model - consult AI Gateway documentation."""
 
@@ -371,6 +567,16 @@ class ParametersTypedDict(TypedDict):
         List[CreateAgentRequestModelConfigurationGuardrailsTypedDict]
     ]
     r"""A list of guardrails to apply to the request."""
+    fallbacks: NotRequired[List[FallbacksTypedDict]]
+    r"""Array of fallback models to use if primary model fails"""
+    retry: NotRequired[RetryTypedDict]
+    r"""Retry configuration for the request"""
+    cache: NotRequired[CacheTypedDict]
+    r"""Cache configuration for the request."""
+    load_balancer: NotRequired[LoadBalancerTypedDict]
+    r"""Load balancer configuration for the request."""
+    timeout: NotRequired[TimeoutTypedDict]
+    r"""Timeout configuration to apply to the request. If the request exceeds the timeout, it will be retried or fallback to the next model if configured."""
 
 
 class Parameters(BaseModel):
@@ -452,77 +658,96 @@ class Parameters(BaseModel):
     guardrails: Optional[List[CreateAgentRequestModelConfigurationGuardrails]] = None
     r"""A list of guardrails to apply to the request."""
 
+    fallbacks: Optional[List[Fallbacks]] = None
+    r"""Array of fallback models to use if primary model fails"""
+
+    retry: Optional[Retry] = None
+    r"""Retry configuration for the request"""
+
+    cache: Optional[Cache] = None
+    r"""Cache configuration for the request."""
+
+    load_balancer: Optional[LoadBalancer] = None
+    r"""Load balancer configuration for the request."""
+
+    timeout: Optional[Timeout] = None
+    r"""Timeout configuration to apply to the request. If the request exceeds the timeout, it will be retried or fallback to the next model if configured."""
+
     @model_serializer(mode="wrap")
     def serialize_model(self, handler):
-        optional_fields = [
-            "audio",
-            "frequency_penalty",
-            "max_tokens",
-            "max_completion_tokens",
-            "logprobs",
-            "top_logprobs",
-            "n",
-            "presence_penalty",
-            "response_format",
-            "reasoning_effort",
-            "verbosity",
-            "seed",
-            "stop",
-            "stream_options",
-            "thinking",
-            "temperature",
-            "top_p",
-            "top_k",
-            "tool_choice",
-            "parallel_tool_calls",
-            "modalities",
-            "guardrails",
-        ]
-        nullable_fields = [
-            "audio",
-            "frequency_penalty",
-            "max_tokens",
-            "max_completion_tokens",
-            "logprobs",
-            "top_logprobs",
-            "n",
-            "presence_penalty",
-            "seed",
-            "stop",
-            "stream_options",
-            "temperature",
-            "top_p",
-            "top_k",
-            "modalities",
-        ]
-        null_default_fields = []
-
+        optional_fields = set(
+            [
+                "audio",
+                "frequency_penalty",
+                "max_tokens",
+                "max_completion_tokens",
+                "logprobs",
+                "top_logprobs",
+                "n",
+                "presence_penalty",
+                "response_format",
+                "reasoning_effort",
+                "verbosity",
+                "seed",
+                "stop",
+                "stream_options",
+                "thinking",
+                "temperature",
+                "top_p",
+                "top_k",
+                "tool_choice",
+                "parallel_tool_calls",
+                "modalities",
+                "guardrails",
+                "fallbacks",
+                "retry",
+                "cache",
+                "load_balancer",
+                "timeout",
+            ]
+        )
+        nullable_fields = set(
+            [
+                "audio",
+                "frequency_penalty",
+                "max_tokens",
+                "max_completion_tokens",
+                "logprobs",
+                "top_logprobs",
+                "n",
+                "presence_penalty",
+                "seed",
+                "stop",
+                "stream_options",
+                "temperature",
+                "top_p",
+                "top_k",
+                "modalities",
+            ]
+        )
         serialized = handler(self)
-
         m = {}
 
         for n, f in type(self).model_fields.items():
             k = f.alias or n
             val = serialized.get(k)
-            serialized.pop(k, None)
-
-            optional_nullable = k in optional_fields and k in nullable_fields
-            is_set = (
-                self.__pydantic_fields_set__.intersection({n})
-                or k in null_default_fields
-            )  # pylint: disable=no-member
-
-            if val is not None and val != UNSET_SENTINEL:
-                m[k] = val
-            elif val != UNSET_SENTINEL and (
-                not k in optional_fields or (optional_nullable and is_set)
-            ):
-                m[k] = val
+            is_nullable_and_explicitly_set = (
+                k in nullable_fields
+                and (self.__pydantic_fields_set__.intersection({n}))  # pylint: disable=no-member
+            )
+
+            if val != UNSET_SENTINEL:
+                if (
+                    val is not None
+                    or k not in optional_fields
+                    or is_nullable_and_explicitly_set
+                ):
+                    m[k] = val
 
         return m
 
 
-class RetryTypedDict(TypedDict):
+class ModelConfigurationRetryTypedDict(TypedDict):
     r"""Retry configuration for model requests. Retries are triggered for specific HTTP status codes (e.g., 500, 429, 502, 503, 504). Supports configurable retry count (1-5) and custom status codes."""
 
     count: NotRequired[float]
@@ -531,7 +756,7 @@ class RetryTypedDict(TypedDict):
     r"""HTTP status codes that trigger retry logic"""
 
 
-class Retry(BaseModel):
+class ModelConfigurationRetry(BaseModel):
     r"""Retry configuration for model requests. Retries are triggered for specific HTTP status codes (e.g., 500, 429, 502, 503, 504). Supports configurable retry count (1-5) and custom status codes."""
 
     count: Optional[float] = 3
@@ -540,6 +765,22 @@ class Retry(BaseModel):
     on_codes: Optional[List[float]] = None
     r"""HTTP status codes that trigger retry logic"""
 
+    @model_serializer(mode="wrap")
+    def serialize_model(self, handler):
+        optional_fields = set(["count", "on_codes"])
+        serialized = handler(self)
+        m = {}
+
+        for n, f in type(self).model_fields.items():
+            k = f.alias or n
+            val = serialized.get(k)
+
+            if val != UNSET_SENTINEL:
+                if val is not None or k not in optional_fields:
+                    m[k] = val
+
+        return m
+
 
 class ModelConfiguration2TypedDict(TypedDict):
     r"""
@@ -551,7 +792,7 @@ class ModelConfiguration2TypedDict(TypedDict):
     r"""A model ID string (e.g., `openai/gpt-4o` or `anthropic/claude-haiku-4-5-20251001`). Only models that support tool calling can be used with agents."""
     parameters: NotRequired[ParametersTypedDict]
     r"""Model behavior parameters that control how the model generates responses. Common parameters: `temperature` (0-1, randomness), `max_completion_tokens` (max output length), `top_p` (sampling diversity). Advanced: `frequency_penalty`, `presence_penalty`, `response_format` (JSON/structured), `reasoning_effort`, `seed` (reproducibility). Support varies by model - consult AI Gateway documentation."""
-    retry: NotRequired[RetryTypedDict]
+    retry: NotRequired[ModelConfigurationRetryTypedDict]
     r"""Retry configuration for model requests. Retries are triggered for specific HTTP status codes (e.g., 500, 429, 502, 503, 504). Supports configurable retry count (1-5) and custom status codes."""
 
 
@@ -567,9 +808,25 @@ class ModelConfiguration2(BaseModel):
     parameters: Optional[Parameters] = None
     r"""Model behavior parameters that control how the model generates responses. Common parameters: `temperature` (0-1, randomness), `max_completion_tokens` (max output length), `top_p` (sampling diversity). Advanced: `frequency_penalty`, `presence_penalty`, `response_format` (JSON/structured), `reasoning_effort`, `seed` (reproducibility). Support varies by model - consult AI Gateway documentation."""
 
-    retry: Optional[Retry] = None
+    retry: Optional[ModelConfigurationRetry] = None
     r"""Retry configuration for model requests. Retries are triggered for specific HTTP status codes (e.g., 500, 429, 502, 503, 504). Supports configurable retry count (1-5) and custom status codes."""
 
+    @model_serializer(mode="wrap")
+    def serialize_model(self, handler):
+        optional_fields = set(["parameters", "retry"])
+        serialized = handler(self)
+        m = {}
+
+        for n, f in type(self).model_fields.items():
+            k = f.alias or n
+            val = serialized.get(k)
+
+            if val != UNSET_SENTINEL:
+                if val is not None or k not in optional_fields:
+                    m[k] = val
+
+        return m
+
 
 ModelConfigurationTypedDict = TypeAliasType(
     "ModelConfigurationTypedDict", Union[ModelConfiguration2TypedDict, str]
@@ -652,6 +909,22 @@ class CreateAgentRequestResponseFormatAgentsJSONSchema(BaseModel):
     strict: Optional[bool] = False
     r"""Whether to enable strict schema adherence when generating the output. If set to true, the model will always follow the exact schema defined in the schema field. Only a subset of JSON Schema is supported when strict is true."""
 
+    @model_serializer(mode="wrap")
+    def serialize_model(self, handler):
+        optional_fields = set(["description", "schema", "strict"])
+        serialized = handler(self)
+        m = {}
+
+        for n, f in type(self).model_fields.items():
+            k = f.alias or n
+            val = serialized.get(k)
+
+            if val != UNSET_SENTINEL:
+                if val is not None or k not in optional_fields:
+                    m[k] = val
+
+        return m
+
 
 class CreateAgentRequestResponseFormatJSONSchemaTypedDict(TypedDict):
     r"""
@@ -782,6 +1055,22 @@ class FallbackModelConfigurationStreamOptions(BaseModel):
     include_usage: Optional[bool] = None
     r"""If set, an additional chunk will be streamed before the data: [DONE] message. The usage field on this chunk shows the token usage statistics for the entire request, and the choices field will always be an empty array. All other chunks will also include a usage field, but with a null value."""
 
+    @model_serializer(mode="wrap")
+    def serialize_model(self, handler):
+        optional_fields = set(["include_usage"])
+        serialized = handler(self)
+        m = {}
+
+        for n, f in type(self).model_fields.items():
+            k = f.alias or n
+            val = serialized.get(k)
+
+            if val != UNSET_SENTINEL:
+                if val is not None or k not in optional_fields:
+                    m[k] = val
+
+        return m
+
 
 FallbackModelConfigurationThinkingTypedDict = TypeAliasType(
     "FallbackModelConfigurationThinkingTypedDict",
@@ -824,6 +1113,22 @@ class CreateAgentRequestToolChoice2(BaseModel):
     type: Optional[CreateAgentRequestToolChoiceType] = None
     r"""The type of the tool. Currently, only function is supported."""
 
+    @model_serializer(mode="wrap")
+    def serialize_model(self, handler):
+        optional_fields = set(["type"])
+        serialized = handler(self)
+        m = {}
+
+        for n, f in type(self).model_fields.items():
+            k = f.alias or n
+            val = serialized.get(k)
+
+            if val != UNSET_SENTINEL:
+                if val is not None or k not in optional_fields:
+                    m[k] = val
+
+        return m
+
 
 CreateAgentRequestToolChoice1 = Literal[
     "none",
@@ -890,65 +1195,225 @@ class FallbackModelConfigurationGuardrails(BaseModel):
     r"""Determines whether the guardrail runs on the input (user message) or output (model response)."""
 
 
-class FallbackModelConfigurationParametersTypedDict(TypedDict):
-    r"""Optional model parameters specific to this fallback model. Overrides primary model parameters if this fallback is used."""
-
-    audio: NotRequired[Nullable[FallbackModelConfigurationAudioTypedDict]]
-    r"""Parameters for audio output. Required when audio output is requested with modalities: [\"audio\"]. Learn more."""
-    frequency_penalty: NotRequired[Nullable[float]]
-    r"""Number between -2.0 and 2.0. Positive values penalize new tokens based on their existing frequency in the text so far, decreasing the model's likelihood to repeat the same line verbatim."""
-    max_tokens: NotRequired[Nullable[int]]
-    r"""`[Deprecated]`. The maximum number of tokens that can be generated in the chat completion. This value can be used to control costs for text generated via API.
+class FallbackModelConfigurationFallbacksTypedDict(TypedDict):
+    model: str
+    r"""Fallback model identifier"""
 
-    This value is now `deprecated` in favor of `max_completion_tokens`, and is not compatible with o1 series models.
-    """
-    max_completion_tokens: NotRequired[Nullable[int]]
-    r"""An upper bound for the number of tokens that can be generated for a completion, including visible output tokens and reasoning tokens"""
-    logprobs: NotRequired[Nullable[bool]]
-    r"""Whether to return log probabilities of the output tokens or not. If true, returns the log probabilities of each output token returned in the content of message."""
-    top_logprobs: NotRequired[Nullable[int]]
-    r"""An integer between 0 and 20 specifying the number of most likely tokens to return at each token position, each with an associated log probability. logprobs must be set to true if this parameter is used."""
-    n: NotRequired[Nullable[int]]
-    r"""How many chat completion choices to generate for each input message. Note that you will be charged based on the number of generated tokens across all of the choices. Keep n as 1 to minimize costs."""
-    presence_penalty: NotRequired[Nullable[float]]
-    r"""Number between -2.0 and 2.0. Positive values penalize new tokens based on whether they appear in the text so far, increasing the model's likelihood to talk about new topics."""
-    response_format: NotRequired[FallbackModelConfigurationResponseFormatTypedDict]
-    r"""An object specifying the format that the model must output"""
-    reasoning_effort: NotRequired[FallbackModelConfigurationReasoningEffort]
-    r"""Constrains effort on reasoning for [reasoning models](https://platform.openai.com/docs/guides/reasoning). Currently supported values are `none`, `minimal`, `low`, `medium`, `high`, and `xhigh`. Reducing reasoning effort can result in faster responses and fewer tokens used on reasoning in a response.
 
-    - `gpt-5.1` defaults to `none`, which does not perform reasoning. The supported reasoning values for `gpt-5.1` are `none`, `low`, `medium`, and `high`. Tool calls are supported for all reasoning values in gpt-5.1.
-    - All models before `gpt-5.1` default to `medium` reasoning effort, and do not support `none`.
-    - The `gpt-5-pro` model defaults to (and only supports) `high` reasoning effort.
-    - `xhigh` is currently only supported for `gpt-5.1-codex-max`.
+class FallbackModelConfigurationFallbacks(BaseModel):
+    model: str
+    r"""Fallback model identifier"""
 
-    Any of \"none\", \"minimal\", \"low\", \"medium\", \"high\", \"xhigh\".
-    """
-    verbosity: NotRequired[str]
-    r"""Adjusts response verbosity. Lower levels yield shorter answers."""
-    seed: NotRequired[Nullable[float]]
-    r"""If specified, our system will make a best effort to sample deterministically, such that repeated requests with the same seed and parameters should return the same result."""
-    stop: NotRequired[Nullable[FallbackModelConfigurationStopTypedDict]]
-    r"""Up to 4 sequences where the API will stop generating further tokens."""
-    stream_options: NotRequired[
-        Nullable[FallbackModelConfigurationStreamOptionsTypedDict]
-    ]
-    r"""Options for streaming response. Only set this when you set stream: true."""
-    thinking: NotRequired[FallbackModelConfigurationThinkingTypedDict]
-    temperature: NotRequired[Nullable[float]]
-    r"""What sampling temperature to use, between 0 and 2. Higher values like 0.8 will make the output more random, while lower values like 0.2 will make it more focused and deterministic."""
-    top_p: NotRequired[Nullable[float]]
-    r"""An alternative to sampling with temperature, called nucleus sampling, where the model considers the results of the tokens with top_p probability mass."""
-    top_k: NotRequired[Nullable[float]]
-    r"""Limits the model to consider only the top k most likely tokens at each step."""
-    tool_choice: NotRequired[FallbackModelConfigurationToolChoiceTypedDict]
-    r"""Controls which (if any) tool is called by the model."""
-    parallel_tool_calls: NotRequired[bool]
-    r"""Whether to enable parallel function calling during tool use."""
-    modalities: NotRequired[Nullable[List[FallbackModelConfigurationModalities]]]
-    r"""Output types that you would like the model to generate. Most models are capable of generating text, which is the default: [\"text\"]. The gpt-4o-audio-preview model can also be used to generate audio. To request that this model generate both text and audio responses, you can use: [\"text\", \"audio\"]."""
-    guardrails: NotRequired[List[FallbackModelConfigurationGuardrailsTypedDict]]
-    r"""A list of guardrails to apply to the request."""
+
+class FallbackModelConfigurationRetryTypedDict(TypedDict):
+    r"""Retry configuration for the request"""
+
+    count: NotRequired[float]
+    r"""Number of retry attempts (1-5)"""
+    on_codes: NotRequired[List[float]]
+    r"""HTTP status codes that trigger retry logic"""
+
+
+class FallbackModelConfigurationRetry(BaseModel):
+    r"""Retry configuration for the request"""
+
+    count: Optional[float] = 3
+    r"""Number of retry attempts (1-5)"""
+
+    on_codes: Optional[List[float]] = None
+    r"""HTTP status codes that trigger retry logic"""
+
+    @model_serializer(mode="wrap")
+    def serialize_model(self, handler):
+        optional_fields = set(["count", "on_codes"])
+        serialized = handler(self)
+        m = {}
+
+        for n, f in type(self).model_fields.items():
+            k = f.alias or n
+            val = serialized.get(k)
+
+            if val != UNSET_SENTINEL:
+                if val is not None or k not in optional_fields:
+                    m[k] = val
+
+        return m
+
+
+FallbackModelConfigurationType = Literal["exact_match",]
+
+
+class FallbackModelConfigurationCacheTypedDict(TypedDict):
+    r"""Cache configuration for the request."""
+
+    type: FallbackModelConfigurationType
+    ttl: NotRequired[float]
+    r"""Time to live for cached responses in seconds. Maximum 259200 seconds (3 days)."""
+
+
+class FallbackModelConfigurationCache(BaseModel):
+    r"""Cache configuration for the request."""
+
+    type: FallbackModelConfigurationType
+
+    ttl: Optional[float] = 1800
+    r"""Time to live for cached responses in seconds. Maximum 259200 seconds (3 days)."""
+
+    @model_serializer(mode="wrap")
+    def serialize_model(self, handler):
+        optional_fields = set(["ttl"])
+        serialized = handler(self)
+        m = {}
+
+        for n, f in type(self).model_fields.items():
+            k = f.alias or n
+            val = serialized.get(k)
+
+            if val != UNSET_SENTINEL:
+                if val is not None or k not in optional_fields:
+                    m[k] = val
+
+        return m
+
+
+CreateAgentRequestLoadBalancerType = Literal["weight_based",]
+
+
+class LoadBalancerModelsTypedDict(TypedDict):
+    model: str
+    r"""Model identifier for load balancing"""
+    weight: NotRequired[float]
+    r"""Weight assigned to this model for load balancing"""
+
+
+class LoadBalancerModels(BaseModel):
+    model: str
+    r"""Model identifier for load balancing"""
+
+    weight: Optional[float] = 0.5
+    r"""Weight assigned to this model for load balancing"""
+
+    @model_serializer(mode="wrap")
+    def serialize_model(self, handler):
+        optional_fields = set(["weight"])
+        serialized = handler(self)
+        m = {}
+
+        for n, f in type(self).model_fields.items():
+            k = f.alias or n
+            val = serialized.get(k)
+
+            if val != UNSET_SENTINEL:
+                if val is not None or k not in optional_fields:
+                    m[k] = val
+
+        return m
+
+
+class CreateAgentRequestLoadBalancer1TypedDict(TypedDict):
+    type: CreateAgentRequestLoadBalancerType
+    models: List[LoadBalancerModelsTypedDict]
+
+
+class CreateAgentRequestLoadBalancer1(BaseModel):
+    type: CreateAgentRequestLoadBalancerType
+
+    models: List[LoadBalancerModels]
+
+
+FallbackModelConfigurationLoadBalancerTypedDict = (
+    CreateAgentRequestLoadBalancer1TypedDict
+)
+r"""Load balancer configuration for the request."""
+
+
+FallbackModelConfigurationLoadBalancer = CreateAgentRequestLoadBalancer1
+r"""Load balancer configuration for the request."""
+
+
+class FallbackModelConfigurationTimeoutTypedDict(TypedDict):
+    r"""Timeout configuration to apply to the request. If the request exceeds the timeout, it will be retried or fallback to the next model if configured."""
+
+    call_timeout: float
+    r"""Timeout value in milliseconds"""
+
+
+class FallbackModelConfigurationTimeout(BaseModel):
+    r"""Timeout configuration to apply to the request. If the request exceeds the timeout, it will be retried or fallback to the next model if configured."""
+
+    call_timeout: float
+    r"""Timeout value in milliseconds"""
+
+
+class FallbackModelConfigurationParametersTypedDict(TypedDict):
+    r"""Optional model parameters specific to this fallback model. Overrides primary model parameters if this fallback is used."""
+
+    audio: NotRequired[Nullable[FallbackModelConfigurationAudioTypedDict]]
+    r"""Parameters for audio output. Required when audio output is requested with modalities: [\"audio\"]. Learn more."""
+    frequency_penalty: NotRequired[Nullable[float]]
+    r"""Number between -2.0 and 2.0. Positive values penalize new tokens based on their existing frequency in the text so far, decreasing the model's likelihood to repeat the same line verbatim."""
+    max_tokens: NotRequired[Nullable[int]]
+    r"""`[Deprecated]`. The maximum number of tokens that can be generated in the chat completion. This value can be used to control costs for text generated via API.
+
+    This value is now `deprecated` in favor of `max_completion_tokens`, and is not compatible with o1 series models.
+    """
+    max_completion_tokens: NotRequired[Nullable[int]]
+    r"""An upper bound for the number of tokens that can be generated for a completion, including visible output tokens and reasoning tokens"""
+    logprobs: NotRequired[Nullable[bool]]
+    r"""Whether to return log probabilities of the output tokens or not. If true, returns the log probabilities of each output token returned in the content of message."""
+    top_logprobs: NotRequired[Nullable[int]]
+    r"""An integer between 0 and 20 specifying the number of most likely tokens to return at each token position, each with an associated log probability. logprobs must be set to true if this parameter is used."""
+    n: NotRequired[Nullable[int]]
+    r"""How many chat completion choices to generate for each input message. Note that you will be charged based on the number of generated tokens across all of the choices. Keep n as 1 to minimize costs."""
+    presence_penalty: NotRequired[Nullable[float]]
+    r"""Number between -2.0 and 2.0. Positive values penalize new tokens based on whether they appear in the text so far, increasing the model's likelihood to talk about new topics."""
+    response_format: NotRequired[FallbackModelConfigurationResponseFormatTypedDict]
+    r"""An object specifying the format that the model must output"""
+    reasoning_effort: NotRequired[FallbackModelConfigurationReasoningEffort]
+    r"""Constrains effort on reasoning for [reasoning models](https://platform.openai.com/docs/guides/reasoning). Currently supported values are `none`, `minimal`, `low`, `medium`, `high`, and `xhigh`. Reducing reasoning effort can result in faster responses and fewer tokens used on reasoning in a response.
+
+    - `gpt-5.1` defaults to `none`, which does not perform reasoning. The supported reasoning values for `gpt-5.1` are `none`, `low`, `medium`, and `high`. Tool calls are supported for all reasoning values in gpt-5.1.
+    - All models before `gpt-5.1` default to `medium` reasoning effort, and do not support `none`.
+    - The `gpt-5-pro` model defaults to (and only supports) `high` reasoning effort.
+    - `xhigh` is currently only supported for `gpt-5.1-codex-max`.
+
+    Any of \"none\", \"minimal\", \"low\", \"medium\", \"high\", \"xhigh\".
+    """
+    verbosity: NotRequired[str]
+    r"""Adjusts response verbosity. Lower levels yield shorter answers."""
+    seed: NotRequired[Nullable[float]]
+    r"""If specified, our system will make a best effort to sample deterministically, such that repeated requests with the same seed and parameters should return the same result."""
+    stop: NotRequired[Nullable[FallbackModelConfigurationStopTypedDict]]
+    r"""Up to 4 sequences where the API will stop generating further tokens."""
+    stream_options: NotRequired[
+        Nullable[FallbackModelConfigurationStreamOptionsTypedDict]
+    ]
+    r"""Options for streaming response. Only set this when you set stream: true."""
+    thinking: NotRequired[FallbackModelConfigurationThinkingTypedDict]
+    temperature: NotRequired[Nullable[float]]
+    r"""What sampling temperature to use, between 0 and 2. Higher values like 0.8 will make the output more random, while lower values like 0.2 will make it more focused and deterministic."""
+    top_p: NotRequired[Nullable[float]]
+    r"""An alternative to sampling with temperature, called nucleus sampling, where the model considers the results of the tokens with top_p probability mass."""
+    top_k: NotRequired[Nullable[float]]
+    r"""Limits the model to consider only the top k most likely tokens at each step."""
+    tool_choice: NotRequired[FallbackModelConfigurationToolChoiceTypedDict]
+    r"""Controls which (if any) tool is called by the model."""
+    parallel_tool_calls: NotRequired[bool]
+    r"""Whether to enable parallel function calling during tool use."""
+    modalities: NotRequired[Nullable[List[FallbackModelConfigurationModalities]]]
+    r"""Output types that you would like the model to generate. Most models are capable of generating text, which is the default: [\"text\"]. The gpt-4o-audio-preview model can also be used to generate audio. To request that this model generate both text and audio responses, you can use: [\"text\", \"audio\"]."""
+    guardrails: NotRequired[List[FallbackModelConfigurationGuardrailsTypedDict]]
+    r"""A list of guardrails to apply to the request."""
+    fallbacks: NotRequired[List[FallbackModelConfigurationFallbacksTypedDict]]
+    r"""Array of fallback models to use if primary model fails"""
+    retry: NotRequired[FallbackModelConfigurationRetryTypedDict]
+    r"""Retry configuration for the request"""
+    cache: NotRequired[FallbackModelConfigurationCacheTypedDict]
+    r"""Cache configuration for the request."""
+    load_balancer: NotRequired[FallbackModelConfigurationLoadBalancerTypedDict]
+    r"""Load balancer configuration for the request."""
+    timeout: NotRequired[FallbackModelConfigurationTimeoutTypedDict]
+    r"""Timeout configuration to apply to the request. If the request exceeds the timeout, it will be retried or fallback to the next model if configured."""
 
 
 class FallbackModelConfigurationParameters(BaseModel):
@@ -1030,77 +1495,96 @@ class FallbackModelConfigurationParameters(BaseModel):
     guardrails: Optional[List[FallbackModelConfigurationGuardrails]] = None
     r"""A list of guardrails to apply to the request."""
 
+    fallbacks: Optional[List[FallbackModelConfigurationFallbacks]] = None
+    r"""Array of fallback models to use if primary model fails"""
+
+    retry: Optional[FallbackModelConfigurationRetry] = None
+    r"""Retry configuration for the request"""
+
+    cache: Optional[FallbackModelConfigurationCache] = None
+    r"""Cache configuration for the request."""
+
+    load_balancer: Optional[FallbackModelConfigurationLoadBalancer] = None
+    r"""Load balancer configuration for the request."""
+
+    timeout: Optional[FallbackModelConfigurationTimeout] = None
+    r"""Timeout configuration to apply to the request. If the request exceeds the timeout, it will be retried or fallback to the next model if configured."""
+
     @model_serializer(mode="wrap")
     def serialize_model(self, handler):
-        optional_fields = [
-            "audio",
-            "frequency_penalty",
-            "max_tokens",
-            "max_completion_tokens",
-            "logprobs",
-            "top_logprobs",
-            "n",
-            "presence_penalty",
-            "response_format",
-            "reasoning_effort",
-            "verbosity",
-            "seed",
-            "stop",
-            "stream_options",
-            "thinking",
-            "temperature",
-            "top_p",
-            "top_k",
-            "tool_choice",
-            "parallel_tool_calls",
-            "modalities",
-            "guardrails",
-        ]
-        nullable_fields = [
-            "audio",
-            "frequency_penalty",
-            "max_tokens",
-            "max_completion_tokens",
-            "logprobs",
-            "top_logprobs",
-            "n",
-            "presence_penalty",
-            "seed",
-            "stop",
-            "stream_options",
-            "temperature",
-            "top_p",
-            "top_k",
-            "modalities",
-        ]
-        null_default_fields = []
-
+        optional_fields = set(
+            [
+                "audio",
+                "frequency_penalty",
+                "max_tokens",
+                "max_completion_tokens",
+                "logprobs",
+                "top_logprobs",
+                "n",
+                "presence_penalty",
+                "response_format",
+                "reasoning_effort",
+                "verbosity",
+                "seed",
+                "stop",
+                "stream_options",
+                "thinking",
+                "temperature",
+                "top_p",
+                "top_k",
+                "tool_choice",
+                "parallel_tool_calls",
+                "modalities",
+                "guardrails",
+                "fallbacks",
+                "retry",
+                "cache",
+                "load_balancer",
+                "timeout",
+            ]
+        )
+        nullable_fields = set(
+            [
+                "audio",
+                "frequency_penalty",
+                "max_tokens",
+                "max_completion_tokens",
+                "logprobs",
+                "top_logprobs",
+                "n",
+                "presence_penalty",
+                "seed",
+                "stop",
+                "stream_options",
+                "temperature",
+                "top_p",
+                "top_k",
+                "modalities",
+            ]
+        )
         serialized = handler(self)
-
         m = {}
 
         for n, f in type(self).model_fields.items():
             k = f.alias or n
             val = serialized.get(k)
-            serialized.pop(k, None)
-
-            optional_nullable = k in optional_fields and k in nullable_fields
-            is_set = (
-                self.__pydantic_fields_set__.intersection({n})
-                or k in null_default_fields
-            )  # pylint: disable=no-member
-
-            if val is not None and val != UNSET_SENTINEL:
-                m[k] = val
-            elif val != UNSET_SENTINEL and (
-                not k in optional_fields or (optional_nullable and is_set)
-            ):
-                m[k] = val
+            is_nullable_and_explicitly_set = (
+                k in nullable_fields
+                and (self.__pydantic_fields_set__.intersection({n}))  # pylint: disable=no-member
+            )
+
+            if val != UNSET_SENTINEL:
+                if (
+                    val is not None
+                    or k not in optional_fields
+                    or is_nullable_and_explicitly_set
+                ):
+                    m[k] = val
 
         return m
 
 
-class FallbackModelConfigurationRetryTypedDict(TypedDict):
+class CreateAgentRequestFallbackModelConfigurationRetryTypedDict(TypedDict):
     r"""Retry configuration for this fallback model. Allows customizing retry count (1-5) and HTTP status codes that trigger retries."""
 
     count: NotRequired[float]
@@ -1109,7 +1593,7 @@ class FallbackModelConfigurationRetryTypedDict(TypedDict):
     r"""HTTP status codes that trigger retry logic"""
 
 
-class FallbackModelConfigurationRetry(BaseModel):
+class CreateAgentRequestFallbackModelConfigurationRetry(BaseModel):
     r"""Retry configuration for this fallback model. Allows customizing retry count (1-5) and HTTP status codes that trigger retries."""
 
     count: Optional[float] = 3
@@ -1118,6 +1602,22 @@ class FallbackModelConfigurationRetry(BaseModel):
     on_codes: Optional[List[float]] = None
     r"""HTTP status codes that trigger retry logic"""
 
+    @model_serializer(mode="wrap")
+    def serialize_model(self, handler):
+        optional_fields = set(["count", "on_codes"])
+        serialized = handler(self)
+        m = {}
+
+        for n, f in type(self).model_fields.items():
+            k = f.alias or n
+            val = serialized.get(k)
+
+            if val != UNSET_SENTINEL:
+                if val is not None or k not in optional_fields:
+                    m[k] = val
+
+        return m
+
 
 class FallbackModelConfiguration2TypedDict(TypedDict):
     r"""Fallback model configuration with optional parameters and retry settings."""
@@ -1126,7 +1626,7 @@ class FallbackModelConfiguration2TypedDict(TypedDict):
     r"""A fallback model ID string. Must support tool calling."""
     parameters: NotRequired[FallbackModelConfigurationParametersTypedDict]
     r"""Optional model parameters specific to this fallback model. Overrides primary model parameters if this fallback is used."""
-    retry: NotRequired[FallbackModelConfigurationRetryTypedDict]
+    retry: NotRequired[CreateAgentRequestFallbackModelConfigurationRetryTypedDict]
     r"""Retry configuration for this fallback model. Allows customizing retry count (1-5) and HTTP status codes that trigger retries."""
 
 
@@ -1139,9 +1639,25 @@ class FallbackModelConfiguration2(BaseModel):
     parameters: Optional[FallbackModelConfigurationParameters] = None
     r"""Optional model parameters specific to this fallback model. Overrides primary model parameters if this fallback is used."""
 
-    retry: Optional[FallbackModelConfigurationRetry] = None
+    retry: Optional[CreateAgentRequestFallbackModelConfigurationRetry] = None
     r"""Retry configuration for this fallback model. Allows customizing retry count (1-5) and HTTP status codes that trigger retries."""
 
+    @model_serializer(mode="wrap")
+    def serialize_model(self, handler):
+        optional_fields = set(["parameters", "retry"])
+        serialized = handler(self)
+        m = {}
+
+        for n, f in type(self).model_fields.items():
+            k = f.alias or n
+            val = serialized.get(k)
+
+            if val != UNSET_SENTINEL:
+                if val is not None or k not in optional_fields:
+                    m[k] = val
+
+        return m
+
 
 FallbackModelConfigurationTypedDict = TypeAliasType(
     "FallbackModelConfigurationTypedDict",
@@ -1164,7 +1680,7 @@ CreateAgentRequestToolApprovalRequired = Literal[
 r"""If all, the agent will require approval for all tools. If respect_tool, the agent will require approval for tools that have the requires_approval flag set to true. If none, the agent will not require approval for any tools."""
 
 
-CreateAgentRequestAgentToolInputCRUDAgentsRequestRequestBodySettingsTools15Type = (
+CreateAgentRequestAgentToolInputCRUDAgentsRequestRequestBodySettingsTools16Type = (
     Literal["mcp",]
 )
 r"""MCP tool type"""
@@ -1176,7 +1692,7 @@ class MCPToolTypedDict(TypedDict):
     tool_id: str
     r"""The ID of the specific nested tool within the MCP server"""
     type: NotRequired[
-        CreateAgentRequestAgentToolInputCRUDAgentsRequestRequestBodySettingsTools15Type
+        CreateAgentRequestAgentToolInputCRUDAgentsRequestRequestBodySettingsTools16Type
     ]
     r"""MCP tool type"""
     key: NotRequired[str]
@@ -1194,7 +1710,7 @@ class MCPTool(BaseModel):
     r"""The ID of the specific nested tool within the MCP server"""
 
     type: Optional[
-        CreateAgentRequestAgentToolInputCRUDAgentsRequestRequestBodySettingsTools15Type
+        CreateAgentRequestAgentToolInputCRUDAgentsRequestRequestBodySettingsTools16Type
     ] = "mcp"
     r"""MCP tool type"""
 
@@ -1207,6 +1723,77 @@ class MCPTool(BaseModel):
     requires_approval: Optional[bool] = False
     r"""Whether this tool requires approval before execution"""
 
+    @model_serializer(mode="wrap")
+    def serialize_model(self, handler):
+        optional_fields = set(["type", "key", "id", "requires_approval"])
+        serialized = handler(self)
+        m = {}
+
+        for n, f in type(self).model_fields.items():
+            k = f.alias or n
+            val = serialized.get(k)
+
+            if val != UNSET_SENTINEL:
+                if val is not None or k not in optional_fields:
+                    m[k] = val
+
+        return m
+
+
+CreateAgentRequestAgentToolInputCRUDAgentsRequestRequestBodySettingsTools15Type = (
+    Literal["json_schema",]
+)
+r"""JSON Schema tool type"""
+
+
+class JSONSchemaToolTypedDict(TypedDict):
+    r"""Enforces structured output format using JSON Schema. Must reference a pre-created JSON Schema tool by key or id."""
+
+    type: NotRequired[
+        CreateAgentRequestAgentToolInputCRUDAgentsRequestRequestBodySettingsTools15Type
+    ]
+    r"""JSON Schema tool type"""
+    key: NotRequired[str]
+    r"""The key of the pre-created JSON Schema tool"""
+    id: NotRequired[str]
+    r"""The ID of the pre-created JSON Schema tool"""
+    requires_approval: NotRequired[bool]
+    r"""Whether this tool requires approval before execution"""
+
+
+class JSONSchemaTool(BaseModel):
+    r"""Enforces structured output format using JSON Schema. Must reference a pre-created JSON Schema tool by key or id."""
+
+    type: Optional[
+        CreateAgentRequestAgentToolInputCRUDAgentsRequestRequestBodySettingsTools15Type
+    ] = "json_schema"
+    r"""JSON Schema tool type"""
+
+    key: Optional[str] = None
+    r"""The key of the pre-created JSON Schema tool"""
+
+    id: Optional[str] = None
+    r"""The ID of the pre-created JSON Schema tool"""
+
+    requires_approval: Optional[bool] = False
+    r"""Whether this tool requires approval before execution"""
+
+    @model_serializer(mode="wrap")
+    def serialize_model(self, handler):
+        optional_fields = set(["type", "key", "id", "requires_approval"])
+        serialized = handler(self)
+        m = {}
+
+        for n, f in type(self).model_fields.items():
+            k = f.alias or n
+            val = serialized.get(k)
+
+            if val != UNSET_SENTINEL:
+                if val is not None or k not in optional_fields:
+                    m[k] = val
+
+        return m
+
 
 CreateAgentRequestAgentToolInputCRUDAgentsRequestRequestBodySettingsTools14Type = (
     Literal["function",]
@@ -1246,6 +1833,22 @@ class FunctionTool(BaseModel):
     requires_approval: Optional[bool] = False
     r"""Whether this tool requires approval before execution"""
 
+    @model_serializer(mode="wrap")
+    def serialize_model(self, handler):
+        optional_fields = set(["type", "key", "id", "requires_approval"])
+        serialized = handler(self)
+        m = {}
+
+        for n, f in type(self).model_fields.items():
+            k = f.alias or n
+            val = serialized.get(k)
+
+            if val != UNSET_SENTINEL:
+                if val is not None or k not in optional_fields:
+                    m[k] = val
+
+        return m
+
 
 CreateAgentRequestAgentToolInputCRUDAgentsRequestRequestBodySettingsTools13Type = (
     Literal["code",]
@@ -1285,6 +1888,22 @@ class CodeExecutionTool(BaseModel):
     requires_approval: Optional[bool] = False
     r"""Whether this tool requires approval before execution"""
 
+    @model_serializer(mode="wrap")
+    def serialize_model(self, handler):
+        optional_fields = set(["type", "key", "id", "requires_approval"])
+        serialized = handler(self)
+        m = {}
+
+        for n, f in type(self).model_fields.items():
+            k = f.alias or n
+            val = serialized.get(k)
+
+            if val != UNSET_SENTINEL:
+                if val is not None or k not in optional_fields:
+                    m[k] = val
+
+        return m
+
 
 CreateAgentRequestAgentToolInputCRUDAgentsRequestRequestBodySettingsTools12Type = (
     Literal["http",]
@@ -1324,6 +1943,22 @@ class HTTPTool(BaseModel):
     requires_approval: Optional[bool] = False
     r"""Whether this tool requires approval before execution"""
 
+    @model_serializer(mode="wrap")
+    def serialize_model(self, handler):
+        optional_fields = set(["type", "key", "id", "requires_approval"])
+        serialized = handler(self)
+        m = {}
+
+        for n, f in type(self).model_fields.items():
+            k = f.alias or n
+            val = serialized.get(k)
+
+            if val != UNSET_SENTINEL:
+                if val is not None or k not in optional_fields:
+                    m[k] = val
+
+        return m
+
 
 CreateAgentRequestAgentToolInputCRUDAgentsRequestRequestBodySettingsTools11Type = (
     Literal["current_date",]
@@ -1350,6 +1985,22 @@ class CurrentDateTool(BaseModel):
     requires_approval: Optional[bool] = None
     r"""Whether this tool requires approval before execution"""
 
+    @model_serializer(mode="wrap")
+    def serialize_model(self, handler):
+        optional_fields = set(["requires_approval"])
+        serialized = handler(self)
+        m = {}
+
+        for n, f in type(self).model_fields.items():
+            k = f.alias or n
+            val = serialized.get(k)
+
+            if val != UNSET_SENTINEL:
+                if val is not None or k not in optional_fields:
+                    m[k] = val
+
+        return m
+
 
 CreateAgentRequestAgentToolInputCRUDAgentsRequestRequestBodySettingsTools10Type = (
     Literal["query_knowledge_base",]
@@ -1376,6 +2027,22 @@ class QueryKnowledgeBaseTool(BaseModel):
     requires_approval: Optional[bool] = None
     r"""Whether this tool requires approval before execution"""
 
+    @model_serializer(mode="wrap")
+    def serialize_model(self, handler):
+        optional_fields = set(["requires_approval"])
+        serialized = handler(self)
+        m = {}
+
+        for n, f in type(self).model_fields.items():
+            k = f.alias or n
+            val = serialized.get(k)
+
+            if val != UNSET_SENTINEL:
+                if val is not None or k not in optional_fields:
+                    m[k] = val
+
+        return m
+
 
 CreateAgentRequestAgentToolInputCRUDAgentsRequestRequestBodySettingsTools9Type = (
     Literal["retrieve_knowledge_bases",]
@@ -1398,6 +2065,22 @@ class RetrieveKnowledgeBasesTool(BaseModel):
     requires_approval: Optional[bool] = None
     r"""Whether this tool requires approval before execution"""
 
+    @model_serializer(mode="wrap")
+    def serialize_model(self, handler):
+        optional_fields = set(["requires_approval"])
+        serialized = handler(self)
+        m = {}
+
+        for n, f in type(self).model_fields.items():
+            k = f.alias or n
+            val = serialized.get(k)
+
+            if val != UNSET_SENTINEL:
+                if val is not None or k not in optional_fields:
+                    m[k] = val
+
+        return m
+
 
 CreateAgentRequestAgentToolInputCRUDAgentsRequestRequestBodySettingsTools8Type = (
     Literal["delete_memory_document",]
@@ -1420,6 +2103,22 @@ class DeleteMemoryDocumentTool(BaseModel):
     requires_approval: Optional[bool] = None
     r"""Whether this tool requires approval before execution"""
 
+    @model_serializer(mode="wrap")
+    def serialize_model(self, handler):
+        optional_fields = set(["requires_approval"])
+        serialized = handler(self)
+        m = {}
+
+        for n, f in type(self).model_fields.items():
+            k = f.alias or n
+            val = serialized.get(k)
+
+            if val != UNSET_SENTINEL:
+                if val is not None or k not in optional_fields:
+                    m[k] = val
+
+        return m
+
 
 CreateAgentRequestAgentToolInputCRUDAgentsRequestRequestBodySettingsToolsType = Literal[
     "retrieve_memory_stores",
@@ -1442,6 +2141,22 @@ class RetrieveMemoryStoresTool(BaseModel):
     requires_approval: Optional[bool] = None
     r"""Whether this tool requires approval before execution"""
 
+    @model_serializer(mode="wrap")
+    def serialize_model(self, handler):
+        optional_fields = set(["requires_approval"])
+        serialized = handler(self)
+        m = {}
+
+        for n, f in type(self).model_fields.items():
+            k = f.alias or n
+            val = serialized.get(k)
+
+            if val != UNSET_SENTINEL:
+                if val is not None or k not in optional_fields:
+                    m[k] = val
+
+        return m
+
 
 CreateAgentRequestAgentToolInputCRUDAgentsRequestRequestBodySettingsType = Literal[
     "write_memory_store",
@@ -1464,6 +2179,22 @@ class WriteMemoryStoreTool(BaseModel):
     requires_approval: Optional[bool] = None
     r"""Whether this tool requires approval before execution"""
 
+    @model_serializer(mode="wrap")
+    def serialize_model(self, handler):
+        optional_fields = set(["requires_approval"])
+        serialized = handler(self)
+        m = {}
+
+        for n, f in type(self).model_fields.items():
+            k = f.alias or n
+            val = serialized.get(k)
+
+            if val != UNSET_SENTINEL:
+                if val is not None or k not in optional_fields:
+                    m[k] = val
+
+        return m
+
 
 CreateAgentRequestAgentToolInputCRUDAgentsRequestRequestBodyType = Literal[
     "query_memory_store",
@@ -1486,6 +2217,22 @@ class QueryMemoryStoreTool(BaseModel):
     requires_approval: Optional[bool] = None
     r"""Whether this tool requires approval before execution"""
 
+    @model_serializer(mode="wrap")
+    def serialize_model(self, handler):
+        optional_fields = set(["requires_approval"])
+        serialized = handler(self)
+        m = {}
+
+        for n, f in type(self).model_fields.items():
+            k = f.alias or n
+            val = serialized.get(k)
+
+            if val != UNSET_SENTINEL:
+                if val is not None or k not in optional_fields:
+                    m[k] = val
+
+        return m
+
 
 CreateAgentRequestAgentToolInputCRUDAgentsRequestType = Literal["retrieve_agents",]
 
@@ -1506,6 +2253,22 @@ class RetrieveAgentsTool(BaseModel):
     requires_approval: Optional[bool] = None
     r"""Whether this tool requires approval before execution"""
 
+    @model_serializer(mode="wrap")
+    def serialize_model(self, handler):
+        optional_fields = set(["requires_approval"])
+        serialized = handler(self)
+        m = {}
+
+        for n, f in type(self).model_fields.items():
+            k = f.alias or n
+            val = serialized.get(k)
+
+            if val != UNSET_SENTINEL:
+                if val is not None or k not in optional_fields:
+                    m[k] = val
+
+        return m
+
 
 CreateAgentRequestAgentToolInputCRUDAgentsType = Literal["call_sub_agent",]
 
@@ -1526,6 +2289,22 @@ class CallSubAgentTool(BaseModel):
     requires_approval: Optional[bool] = None
     r"""Whether this tool requires approval before execution"""
 
+    @model_serializer(mode="wrap")
+    def serialize_model(self, handler):
+        optional_fields = set(["requires_approval"])
+        serialized = handler(self)
+        m = {}
+
+        for n, f in type(self).model_fields.items():
+            k = f.alias or n
+            val = serialized.get(k)
+
+            if val != UNSET_SENTINEL:
+                if val is not None or k not in optional_fields:
+                    m[k] = val
+
+        return m
+
 
 CreateAgentRequestAgentToolInputCRUDType = Literal["web_scraper",]
 
@@ -1546,6 +2325,22 @@ class WebScraperTool(BaseModel):
     requires_approval: Optional[bool] = None
     r"""Whether this tool requires approval before execution"""
 
+    @model_serializer(mode="wrap")
+    def serialize_model(self, handler):
+        optional_fields = set(["requires_approval"])
+        serialized = handler(self)
+        m = {}
+
+        for n, f in type(self).model_fields.items():
+            k = f.alias or n
+            val = serialized.get(k)
+
+            if val != UNSET_SENTINEL:
+                if val is not None or k not in optional_fields:
+                    m[k] = val
+
+        return m
+
 
 AgentToolInputCRUDType = Literal["google_search",]
 
@@ -1566,6 +2361,22 @@ class GoogleSearchTool(BaseModel):
     requires_approval: Optional[bool] = None
     r"""Whether this tool requires approval before execution"""
 
+    @model_serializer(mode="wrap")
+    def serialize_model(self, handler):
+        optional_fields = set(["requires_approval"])
+        serialized = handler(self)
+        m = {}
+
+        for n, f in type(self).model_fields.items():
+            k = f.alias or n
+            val = serialized.get(k)
+
+            if val != UNSET_SENTINEL:
+                if val is not None or k not in optional_fields:
+                    m[k] = val
+
+        return m
+
 
 AgentToolInputCRUDTypedDict = TypeAliasType(
     "AgentToolInputCRUDTypedDict",
@@ -1584,10 +2395,11 @@ AgentToolInputCRUDTypedDict = TypeAliasType(
         HTTPToolTypedDict,
         CodeExecutionToolTypedDict,
         FunctionToolTypedDict,
+        JSONSchemaToolTypedDict,
         MCPToolTypedDict,
     ],
 )
-r"""Tool configuration for agent create/update operations. Built-in tools only require a type, while custom tools (HTTP, Code, Function, MCP) must reference pre-created tools by key or id."""
+r"""Tool configuration for agent create/update operations. Built-in tools only require a type, while custom tools (HTTP, Code, Function, JSON Schema, MCP) must reference pre-created tools by key or id."""
 
 
 AgentToolInputCRUD = TypeAliasType(
@@ -1607,10 +2419,11 @@ AgentToolInputCRUD = TypeAliasType(
         HTTPTool,
         CodeExecutionTool,
         FunctionTool,
+        JSONSchemaTool,
         MCPTool,
     ],
 )
-r"""Tool configuration for agent create/update operations. Built-in tools only require a type, while custom tools (HTTP, Code, Function, MCP) must reference pre-created tools by key or id."""
+r"""Tool configuration for agent create/update operations. Built-in tools only require a type, while custom tools (HTTP, Code, Function, JSON Schema, MCP) must reference pre-created tools by key or id."""
 
 
 CreateAgentRequestExecuteOn = Literal[
@@ -1639,6 +2452,22 @@ class CreateAgentRequestEvaluators(BaseModel):
     sample_rate: Optional[float] = 50
     r"""The percentage of executions to evaluate with this evaluator (1-100). For example, a value of 50 means the evaluator will run on approximately half of the executions."""
 
+    @model_serializer(mode="wrap")
+    def serialize_model(self, handler):
+        optional_fields = set(["sample_rate"])
+        serialized = handler(self)
+        m = {}
+
+        for n, f in type(self).model_fields.items():
+            k = f.alias or n
+            val = serialized.get(k)
+
+            if val != UNSET_SENTINEL:
+                if val is not None or k not in optional_fields:
+                    m[k] = val
+
+        return m
+
 
 CreateAgentRequestAgentsExecuteOn = Literal[
     "input",
@@ -1666,6 +2495,22 @@ class CreateAgentRequestGuardrails(BaseModel):
     sample_rate: Optional[float] = 50
     r"""The percentage of executions to evaluate with this evaluator (1-100). For example, a value of 50 means the evaluator will run on approximately half of the executions."""
 
+    @model_serializer(mode="wrap")
+    def serialize_model(self, handler):
+        optional_fields = set(["sample_rate"])
+        serialized = handler(self)
+        m = {}
+
+        for n, f in type(self).model_fields.items():
+            k = f.alias or n
+            val = serialized.get(k)
+
+            if val != UNSET_SENTINEL:
+                if val is not None or k not in optional_fields:
+                    m[k] = val
+
+        return m
+
 
 class CreateAgentRequestSettingsTypedDict(TypedDict):
     r"""Configuration settings for the agent's behavior"""
@@ -1707,6 +2552,31 @@ class CreateAgentRequestSettings(BaseModel):
     guardrails: Optional[List[CreateAgentRequestGuardrails]] = None
     r"""Configuration for a guardrail applied to the agent"""
 
+    @model_serializer(mode="wrap")
+    def serialize_model(self, handler):
+        optional_fields = set(
+            [
+                "max_iterations",
+                "max_execution_time",
+                "tool_approval_required",
+                "tools",
+                "evaluators",
+                "guardrails",
+            ]
+        )
+        serialized = handler(self)
+        m = {}
+
+        for n, f in type(self).model_fields.items():
+            k = f.alias or n
+            val = serialized.get(k)
+
+            if val != UNSET_SENTINEL:
+                if val is not None or k not in optional_fields:
+                    m[k] = val
+
+        return m
+
 
 class KnowledgeBasesTypedDict(TypedDict):
     knowledge_id: str
@@ -1732,6 +2602,22 @@ class TeamOfAgents(BaseModel):
     role: Optional[str] = None
     r"""The role of the agent in this context. This is used to give extra information to the leader to help it decide which agent to hand off to."""
 
+    @model_serializer(mode="wrap")
+    def serialize_model(self, handler):
+        optional_fields = set(["role"])
+        serialized = handler(self)
+        m = {}
+
+        for n, f in type(self).model_fields.items():
+            k = f.alias or n
+            val = serialized.get(k)
+
+            if val != UNSET_SENTINEL:
+                if val is not None or k not in optional_fields:
+                    m[k] = val
+
+        return m
+
 
 Source = Literal[
     "internal",
@@ -1821,6 +2707,33 @@ class CreateAgentRequestRequestBody(BaseModel):
 
     source: Optional[Source] = None
 
+    @model_serializer(mode="wrap")
+    def serialize_model(self, handler):
+        optional_fields = set(
+            [
+                "display_name",
+                "system_prompt",
+                "fallback_models",
+                "memory_stores",
+                "knowledge_bases",
+                "team_of_agents",
+                "variables",
+                "source",
+            ]
+        )
+        serialized = handler(self)
+        m = {}
+
+        for n, f in type(self).model_fields.items():
+            k = f.alias or n
+            val = serialized.get(k)
+
+            if val != UNSET_SENTINEL:
+                if val is not None or k not in optional_fields:
+                    m[k] = val
+
+        return m
+
 
 CreateAgentRequestStatus = Literal[
     "live",
@@ -1900,6 +2813,32 @@ class CreateAgentRequestTools(BaseModel):
     timeout: Optional[float] = 120
     r"""Tool execution timeout in seconds (default: 2 minutes, max: 10 minutes)"""
 
+    @model_serializer(mode="wrap")
+    def serialize_model(self, handler):
+        optional_fields = set(
+            [
+                "key",
+                "display_name",
+                "description",
+                "requires_approval",
+                "tool_id",
+                "conditions",
+                "timeout",
+            ]
+        )
+        serialized = handler(self)
+        m = {}
+
+        for n, f in type(self).model_fields.items():
+            k = f.alias or n
+            val = serialized.get(k)
+
+            if val != UNSET_SENTINEL:
+                if val is not None or k not in optional_fields:
+                    m[k] = val
+
+        return m
+
 
 CreateAgentRequestAgentsResponseExecuteOn = Literal[
     "input",
@@ -1927,6 +2866,22 @@ class CreateAgentRequestAgentsEvaluators(BaseModel):
     sample_rate: Optional[float] = 50
     r"""The percentage of executions to evaluate with this evaluator (1-100). For example, a value of 50 means the evaluator will run on approximately half of the executions."""
 
+    @model_serializer(mode="wrap")
+    def serialize_model(self, handler):
+        optional_fields = set(["sample_rate"])
+        serialized = handler(self)
+        m = {}
+
+        for n, f in type(self).model_fields.items():
+            k = f.alias or n
+            val = serialized.get(k)
+
+            if val != UNSET_SENTINEL:
+                if val is not None or k not in optional_fields:
+                    m[k] = val
+
+        return m
+
 
 CreateAgentRequestAgentsResponse201ExecuteOn = Literal[
     "input",
@@ -1954,6 +2909,22 @@ class CreateAgentRequestAgentsGuardrails(BaseModel):
     sample_rate: Optional[float] = 50
     r"""The percentage of executions to evaluate with this evaluator (1-100). For example, a value of 50 means the evaluator will run on approximately half of the executions."""
 
+    @model_serializer(mode="wrap")
+    def serialize_model(self, handler):
+        optional_fields = set(["sample_rate"])
+        serialized = handler(self)
+        m = {}
+
+        for n, f in type(self).model_fields.items():
+            k = f.alias or n
+            val = serialized.get(k)
+
+            if val != UNSET_SENTINEL:
+                if val is not None or k not in optional_fields:
+                    m[k] = val
+
+        return m
+
 
 class CreateAgentRequestAgentsSettingsTypedDict(TypedDict):
     max_iterations: NotRequired[int]
@@ -1989,6 +2960,31 @@ class CreateAgentRequestAgentsSettings(BaseModel):
     guardrails: Optional[List[CreateAgentRequestAgentsGuardrails]] = None
     r"""Configuration for a guardrail applied to the agent"""
 
+    @model_serializer(mode="wrap")
+    def serialize_model(self, handler):
+        optional_fields = set(
+            [
+                "max_iterations",
+                "max_execution_time",
+                "tool_approval_required",
+                "tools",
+                "evaluators",
+                "guardrails",
+            ]
+        )
+        serialized = handler(self)
+        m = {}
+
+        for n, f in type(self).model_fields.items():
+            k = f.alias or n
+            val = serialized.get(k)
+
+            if val != UNSET_SENTINEL:
+                if val is not None or k not in optional_fields:
+                    m[k] = val
+
+        return m
+
 
 CreateAgentRequestVoice = Literal[
     "alloy",
@@ -2059,6 +3055,22 @@ class CreateAgentRequestResponseFormatAgentsResponseJSONSchema(BaseModel):
     strict: Optional[bool] = False
     r"""Whether to enable strict schema adherence when generating the output. If set to true, the model will always follow the exact schema defined in the schema field. Only a subset of JSON Schema is supported when strict is true."""
 
+    @model_serializer(mode="wrap")
+    def serialize_model(self, handler):
+        optional_fields = set(["description", "schema", "strict"])
+        serialized = handler(self)
+        m = {}
+
+        for n, f in type(self).model_fields.items():
+            k = f.alias or n
+            val = serialized.get(k)
+
+            if val != UNSET_SENTINEL:
+                if val is not None or k not in optional_fields:
+                    m[k] = val
+
+        return m
+
 
 class CreateAgentRequestResponseFormatAgentsResponse201JSONSchemaTypedDict(TypedDict):
     r"""
@@ -2190,6 +3202,22 @@ class CreateAgentRequestStreamOptions(BaseModel):
     include_usage: Optional[bool] = None
     r"""If set, an additional chunk will be streamed before the data: [DONE] message. The usage field on this chunk shows the token usage statistics for the entire request, and the choices field will always be an empty array. All other chunks will also include a usage field, but with a null value."""
 
+    @model_serializer(mode="wrap")
+    def serialize_model(self, handler):
+        optional_fields = set(["include_usage"])
+        serialized = handler(self)
+        m = {}
+
+        for n, f in type(self).model_fields.items():
+            k = f.alias or n
+            val = serialized.get(k)
+
+            if val != UNSET_SENTINEL:
+                if val is not None or k not in optional_fields:
+                    m[k] = val
+
+        return m
+
 
 CreateAgentRequestThinkingTypedDict = TypeAliasType(
     "CreateAgentRequestThinkingTypedDict",
@@ -2232,6 +3260,22 @@ class CreateAgentRequestToolChoiceAgents2(BaseModel):
     type: Optional[CreateAgentRequestToolChoiceAgentsType] = None
     r"""The type of the tool. Currently, only function is supported."""
 
+    @model_serializer(mode="wrap")
+    def serialize_model(self, handler):
+        optional_fields = set(["type"])
+        serialized = handler(self)
+        m = {}
+
+        for n, f in type(self).model_fields.items():
+            k = f.alias or n
+            val = serialized.get(k)
+
+            if val != UNSET_SENTINEL:
+                if val is not None or k not in optional_fields:
+                    m[k] = val
+
+        return m
+
 
 CreateAgentRequestToolChoiceAgents1 = Literal[
     "none",
@@ -2257,48 +3301,196 @@ CreateAgentRequestToolChoice = TypeAliasType(
 r"""Controls which (if any) tool is called by the model."""
 
 
-CreateAgentRequestModalities = Literal[
-    "text",
-    "audio",
-]
+CreateAgentRequestModalities = Literal[
+    "text",
+    "audio",
+]
+
+
+CreateAgentRequestIDAgents1 = Literal[
+    "orq_pii_detection",
+    "orq_sexual_moderation",
+    "orq_harmful_moderation",
+]
+r"""The key of the guardrail."""
+
+
+CreateAgentRequestIDTypedDict = TypeAliasType(
+    "CreateAgentRequestIDTypedDict", Union[CreateAgentRequestIDAgents1, str]
+)
+
+
+CreateAgentRequestID = TypeAliasType(
+    "CreateAgentRequestID", Union[CreateAgentRequestIDAgents1, str]
+)
+
+
+CreateAgentRequestAgentsResponse201ApplicationJSONExecuteOn = Literal[
+    "input",
+    "output",
+]
+r"""Determines whether the guardrail runs on the input (user message) or output (model response)."""
+
+
+class CreateAgentRequestAgentsResponseGuardrailsTypedDict(TypedDict):
+    id: CreateAgentRequestIDTypedDict
+    execute_on: CreateAgentRequestAgentsResponse201ApplicationJSONExecuteOn
+    r"""Determines whether the guardrail runs on the input (user message) or output (model response)."""
+
+
+class CreateAgentRequestAgentsResponseGuardrails(BaseModel):
+    id: CreateAgentRequestID
+
+    execute_on: CreateAgentRequestAgentsResponse201ApplicationJSONExecuteOn
+    r"""Determines whether the guardrail runs on the input (user message) or output (model response)."""
+
+
+class CreateAgentRequestFallbacksTypedDict(TypedDict):
+    model: str
+    r"""Fallback model identifier"""
+
+
+class CreateAgentRequestFallbacks(BaseModel):
+    model: str
+    r"""Fallback model identifier"""
+
+
+class CreateAgentRequestAgentsRetryTypedDict(TypedDict):
+    r"""Retry configuration for the request"""
+
+    count: NotRequired[float]
+    r"""Number of retry attempts (1-5)"""
+    on_codes: NotRequired[List[float]]
+    r"""HTTP status codes that trigger retry logic"""
+
+
+class CreateAgentRequestAgentsRetry(BaseModel):
+    r"""Retry configuration for the request"""
+
+    count: Optional[float] = 3
+    r"""Number of retry attempts (1-5)"""
+
+    on_codes: Optional[List[float]] = None
+    r"""HTTP status codes that trigger retry logic"""
+
+    @model_serializer(mode="wrap")
+    def serialize_model(self, handler):
+        optional_fields = set(["count", "on_codes"])
+        serialized = handler(self)
+        m = {}
+
+        for n, f in type(self).model_fields.items():
+            k = f.alias or n
+            val = serialized.get(k)
+
+            if val != UNSET_SENTINEL:
+                if val is not None or k not in optional_fields:
+                    m[k] = val
+
+        return m
+
+
+CreateAgentRequestType = Literal["exact_match",]
+
+
+class CreateAgentRequestCacheTypedDict(TypedDict):
+    r"""Cache configuration for the request."""
+
+    type: CreateAgentRequestType
+    ttl: NotRequired[float]
+    r"""Time to live for cached responses in seconds. Maximum 259200 seconds (3 days)."""
+
+
+class CreateAgentRequestCache(BaseModel):
+    r"""Cache configuration for the request."""
+
+    type: CreateAgentRequestType
+
+    ttl: Optional[float] = 1800
+    r"""Time to live for cached responses in seconds. Maximum 259200 seconds (3 days)."""
+
+    @model_serializer(mode="wrap")
+    def serialize_model(self, handler):
+        optional_fields = set(["ttl"])
+        serialized = handler(self)
+        m = {}
+
+        for n, f in type(self).model_fields.items():
+            k = f.alias or n
+            val = serialized.get(k)
+
+            if val != UNSET_SENTINEL:
+                if val is not None or k not in optional_fields:
+                    m[k] = val
+
+        return m
+
+
+CreateAgentRequestLoadBalancerAgentsType = Literal["weight_based",]
+
+
+class CreateAgentRequestLoadBalancerAgentsModelsTypedDict(TypedDict):
+    model: str
+    r"""Model identifier for load balancing"""
+    weight: NotRequired[float]
+    r"""Weight assigned to this model for load balancing"""
+
+
+class CreateAgentRequestLoadBalancerAgentsModels(BaseModel):
+    model: str
+    r"""Model identifier for load balancing"""
+
+    weight: Optional[float] = 0.5
+    r"""Weight assigned to this model for load balancing"""
+
+    @model_serializer(mode="wrap")
+    def serialize_model(self, handler):
+        optional_fields = set(["weight"])
+        serialized = handler(self)
+        m = {}
+
+        for n, f in type(self).model_fields.items():
+            k = f.alias or n
+            val = serialized.get(k)
+
+            if val != UNSET_SENTINEL:
+                if val is not None or k not in optional_fields:
+                    m[k] = val
+
+        return m
 
 
-CreateAgentRequestIDAgents1 = Literal[
-    "orq_pii_detection",
-    "orq_sexual_moderation",
-    "orq_harmful_moderation",
-]
-r"""The key of the guardrail."""
+class CreateAgentRequestLoadBalancerAgents1TypedDict(TypedDict):
+    type: CreateAgentRequestLoadBalancerAgentsType
+    models: List[CreateAgentRequestLoadBalancerAgentsModelsTypedDict]
 
 
-CreateAgentRequestIDTypedDict = TypeAliasType(
-    "CreateAgentRequestIDTypedDict", Union[CreateAgentRequestIDAgents1, str]
-)
+class CreateAgentRequestLoadBalancerAgents1(BaseModel):
+    type: CreateAgentRequestLoadBalancerAgentsType
 
+    models: List[CreateAgentRequestLoadBalancerAgentsModels]
 
-CreateAgentRequestID = TypeAliasType(
-    "CreateAgentRequestID", Union[CreateAgentRequestIDAgents1, str]
-)
 
+CreateAgentRequestLoadBalancerTypedDict = CreateAgentRequestLoadBalancerAgents1TypedDict
+r"""Load balancer configuration for the request."""
 
-CreateAgentRequestAgentsResponse201ApplicationJSONExecuteOn = Literal[
-    "input",
-    "output",
-]
-r"""Determines whether the guardrail runs on the input (user message) or output (model response)."""
 
+CreateAgentRequestLoadBalancer = CreateAgentRequestLoadBalancerAgents1
+r"""Load balancer configuration for the request."""
 
-class CreateAgentRequestAgentsResponseGuardrailsTypedDict(TypedDict):
-    id: CreateAgentRequestIDTypedDict
-    execute_on: CreateAgentRequestAgentsResponse201ApplicationJSONExecuteOn
-    r"""Determines whether the guardrail runs on the input (user message) or output (model response)."""
 
+class CreateAgentRequestTimeoutTypedDict(TypedDict):
+    r"""Timeout configuration to apply to the request. If the request exceeds the timeout, it will be retried or fallback to the next model if configured."""
 
-class CreateAgentRequestAgentsResponseGuardrails(BaseModel):
-    id: CreateAgentRequestID
+    call_timeout: float
+    r"""Timeout value in milliseconds"""
 
-    execute_on: CreateAgentRequestAgentsResponse201ApplicationJSONExecuteOn
-    r"""Determines whether the guardrail runs on the input (user message) or output (model response)."""
+
+class CreateAgentRequestTimeout(BaseModel):
+    r"""Timeout configuration to apply to the request. If the request exceeds the timeout, it will be retried or fallback to the next model if configured."""
+
+    call_timeout: float
+    r"""Timeout value in milliseconds"""
 
 
 class CreateAgentRequestParametersTypedDict(TypedDict):
@@ -2358,6 +3550,16 @@ class CreateAgentRequestParametersTypedDict(TypedDict):
     r"""Output types that you would like the model to generate. Most models are capable of generating text, which is the default: [\"text\"]. The gpt-4o-audio-preview model can also be used to generate audio. To request that this model generate both text and audio responses, you can use: [\"text\", \"audio\"]."""
     guardrails: NotRequired[List[CreateAgentRequestAgentsResponseGuardrailsTypedDict]]
     r"""A list of guardrails to apply to the request."""
+    fallbacks: NotRequired[List[CreateAgentRequestFallbacksTypedDict]]
+    r"""Array of fallback models to use if primary model fails"""
+    retry: NotRequired[CreateAgentRequestAgentsRetryTypedDict]
+    r"""Retry configuration for the request"""
+    cache: NotRequired[CreateAgentRequestCacheTypedDict]
+    r"""Cache configuration for the request."""
+    load_balancer: NotRequired[CreateAgentRequestLoadBalancerTypedDict]
+    r"""Load balancer configuration for the request."""
+    timeout: NotRequired[CreateAgentRequestTimeoutTypedDict]
+    r"""Timeout configuration to apply to the request. If the request exceeds the timeout, it will be retried or fallback to the next model if configured."""
 
 
 class CreateAgentRequestParameters(BaseModel):
@@ -2439,72 +3641,91 @@ class CreateAgentRequestParameters(BaseModel):
     guardrails: Optional[List[CreateAgentRequestAgentsResponseGuardrails]] = None
     r"""A list of guardrails to apply to the request."""
 
+    fallbacks: Optional[List[CreateAgentRequestFallbacks]] = None
+    r"""Array of fallback models to use if primary model fails"""
+
+    retry: Optional[CreateAgentRequestAgentsRetry] = None
+    r"""Retry configuration for the request"""
+
+    cache: Optional[CreateAgentRequestCache] = None
+    r"""Cache configuration for the request."""
+
+    load_balancer: Optional[CreateAgentRequestLoadBalancer] = None
+    r"""Load balancer configuration for the request."""
+
+    timeout: Optional[CreateAgentRequestTimeout] = None
+    r"""Timeout configuration to apply to the request. If the request exceeds the timeout, it will be retried or fallback to the next model if configured."""
+
     @model_serializer(mode="wrap")
     def serialize_model(self, handler):
-        optional_fields = [
-            "audio",
-            "frequency_penalty",
-            "max_tokens",
-            "max_completion_tokens",
-            "logprobs",
-            "top_logprobs",
-            "n",
-            "presence_penalty",
-            "response_format",
-            "reasoning_effort",
-            "verbosity",
-            "seed",
-            "stop",
-            "stream_options",
-            "thinking",
-            "temperature",
-            "top_p",
-            "top_k",
-            "tool_choice",
-            "parallel_tool_calls",
-            "modalities",
-            "guardrails",
-        ]
-        nullable_fields = [
-            "audio",
-            "frequency_penalty",
-            "max_tokens",
-            "max_completion_tokens",
-            "logprobs",
-            "top_logprobs",
-            "n",
-            "presence_penalty",
-            "seed",
-            "stop",
-            "stream_options",
-            "temperature",
-            "top_p",
-            "top_k",
-            "modalities",
-        ]
-        null_default_fields = []
-
+        optional_fields = set(
+            [
+                "audio",
+                "frequency_penalty",
+                "max_tokens",
+                "max_completion_tokens",
+                "logprobs",
+                "top_logprobs",
+                "n",
+                "presence_penalty",
+                "response_format",
+                "reasoning_effort",
+                "verbosity",
+                "seed",
+                "stop",
+                "stream_options",
+                "thinking",
+                "temperature",
+                "top_p",
+                "top_k",
+                "tool_choice",
+                "parallel_tool_calls",
+                "modalities",
+                "guardrails",
+                "fallbacks",
+                "retry",
+                "cache",
+                "load_balancer",
+                "timeout",
+            ]
+        )
+        nullable_fields = set(
+            [
+                "audio",
+                "frequency_penalty",
+                "max_tokens",
+                "max_completion_tokens",
+                "logprobs",
+                "top_logprobs",
+                "n",
+                "presence_penalty",
+                "seed",
+                "stop",
+                "stream_options",
+                "temperature",
+                "top_p",
+                "top_k",
+                "modalities",
+            ]
+        )
         serialized = handler(self)
-
         m = {}
 
         for n, f in type(self).model_fields.items():
             k = f.alias or n
             val = serialized.get(k)
-            serialized.pop(k, None)
-
-            optional_nullable = k in optional_fields and k in nullable_fields
-            is_set = (
-                self.__pydantic_fields_set__.intersection({n})
-                or k in null_default_fields
-            )  # pylint: disable=no-member
-
-            if val is not None and val != UNSET_SENTINEL:
-                m[k] = val
-            elif val != UNSET_SENTINEL and (
-                not k in optional_fields or (optional_nullable and is_set)
-            ):
-                m[k] = val
+            is_nullable_and_explicitly_set = (
+                k in nullable_fields
+                and (self.__pydantic_fields_set__.intersection({n}))  # pylint: disable=no-member
+            )
+
+            if val != UNSET_SENTINEL:
+                if (
+                    val is not None
+                    or k not in optional_fields
+                    or is_nullable_and_explicitly_set
+                ):
+                    m[k] = val
 
         return m
 
@@ -2527,6 +3748,22 @@ class CreateAgentRequestRetry(BaseModel):
     on_codes: Optional[List[float]] = None
     r"""HTTP status codes that trigger retry logic"""
 
+    @model_serializer(mode="wrap")
+    def serialize_model(self, handler):
+        optional_fields = set(["count", "on_codes"])
+        serialized = handler(self)
+        m = {}
+
+        for n, f in type(self).model_fields.items():
+            k = f.alias or n
+            val = serialized.get(k)
+
+            if val != UNSET_SENTINEL:
+                if val is not None or k not in optional_fields:
+                    m[k] = val
+
+        return m
+
 
 CreateAgentRequestFallbackModelConfigurationVoice = Literal[
     "alloy",
@@ -2604,6 +3841,22 @@ class CreateAgentRequestResponseFormatAgentsResponse201ApplicationJSONJSONSchema
     strict: Optional[bool] = False
     r"""Whether to enable strict schema adherence when generating the output. If set to true, the model will always follow the exact schema defined in the schema field. Only a subset of JSON Schema is supported when strict is true."""
 
+    @model_serializer(mode="wrap")
+    def serialize_model(self, handler):
+        optional_fields = set(["description", "schema", "strict"])
+        serialized = handler(self)
+        m = {}
+
+        for n, f in type(self).model_fields.items():
+            k = f.alias or n
+            val = serialized.get(k)
+
+            if val != UNSET_SENTINEL:
+                if val is not None or k not in optional_fields:
+                    m[k] = val
+
+        return m
+
 
 class CreateAgentRequestResponseFormatAgentsResponse201ApplicationJSONResponseBodyJSONSchemaTypedDict(
     TypedDict
@@ -2751,6 +4004,22 @@ class CreateAgentRequestFallbackModelConfigurationStreamOptions(BaseModel):
     include_usage: Optional[bool] = None
     r"""If set, an additional chunk will be streamed before the data: [DONE] message. The usage field on this chunk shows the token usage statistics for the entire request, and the choices field will always be an empty array. All other chunks will also include a usage field, but with a null value."""
 
+    @model_serializer(mode="wrap")
+    def serialize_model(self, handler):
+        optional_fields = set(["include_usage"])
+        serialized = handler(self)
+        m = {}
+
+        for n, f in type(self).model_fields.items():
+            k = f.alias or n
+            val = serialized.get(k)
+
+            if val != UNSET_SENTINEL:
+                if val is not None or k not in optional_fields:
+                    m[k] = val
+
+        return m
+
 
 CreateAgentRequestFallbackModelConfigurationThinkingTypedDict = TypeAliasType(
     "CreateAgentRequestFallbackModelConfigurationThinkingTypedDict",
@@ -2793,6 +4062,22 @@ class CreateAgentRequestToolChoiceAgentsResponse2(BaseModel):
     type: Optional[CreateAgentRequestToolChoiceAgentsResponseType] = None
     r"""The type of the tool. Currently, only function is supported."""
 
+    @model_serializer(mode="wrap")
+    def serialize_model(self, handler):
+        optional_fields = set(["type"])
+        serialized = handler(self)
+        m = {}
+
+        for n, f in type(self).model_fields.items():
+            k = f.alias or n
+            val = serialized.get(k)
+
+            if val != UNSET_SENTINEL:
+                if val is not None or k not in optional_fields:
+                    m[k] = val
+
+        return m
+
 
 CreateAgentRequestToolChoiceAgentsResponse1 = Literal[
     "none",
@@ -2867,6 +4152,160 @@ class CreateAgentRequestFallbackModelConfigurationGuardrails(BaseModel):
     r"""Determines whether the guardrail runs on the input (user message) or output (model response)."""
 
 
+class CreateAgentRequestFallbackModelConfigurationFallbacksTypedDict(TypedDict):
+    model: str
+    r"""Fallback model identifier"""
+
+
+class CreateAgentRequestFallbackModelConfigurationFallbacks(BaseModel):
+    model: str
+    r"""Fallback model identifier"""
+
+
+class CreateAgentRequestFallbackModelConfigurationAgentsResponseRetryTypedDict(
+    TypedDict
+):
+    r"""Retry configuration for the request"""
+
+    count: NotRequired[float]
+    r"""Number of retry attempts (1-5)"""
+    on_codes: NotRequired[List[float]]
+    r"""HTTP status codes that trigger retry logic"""
+
+
+class CreateAgentRequestFallbackModelConfigurationAgentsResponseRetry(BaseModel):
+    r"""Retry configuration for the request"""
+
+    count: Optional[float] = 3
+    r"""Number of retry attempts (1-5)"""
+
+    on_codes: Optional[List[float]] = None
+    r"""HTTP status codes that trigger retry logic"""
+
+    @model_serializer(mode="wrap")
+    def serialize_model(self, handler):
+        optional_fields = set(["count", "on_codes"])
+        serialized = handler(self)
+        m = {}
+
+        for n, f in type(self).model_fields.items():
+            k = f.alias or n
+            val = serialized.get(k)
+
+            if val != UNSET_SENTINEL:
+                if val is not None or k not in optional_fields:
+                    m[k] = val
+
+        return m
+
+
+CreateAgentRequestFallbackModelConfigurationType = Literal["exact_match",]
+
+
+class CreateAgentRequestFallbackModelConfigurationCacheTypedDict(TypedDict):
+    r"""Cache configuration for the request."""
+
+    type: CreateAgentRequestFallbackModelConfigurationType
+    ttl: NotRequired[float]
+    r"""Time to live for cached responses in seconds. Maximum 259200 seconds (3 days)."""
+
+
+class CreateAgentRequestFallbackModelConfigurationCache(BaseModel):
+    r"""Cache configuration for the request."""
+
+    type: CreateAgentRequestFallbackModelConfigurationType
+
+    ttl: Optional[float] = 1800
+    r"""Time to live for cached responses in seconds. Maximum 259200 seconds (3 days)."""
+
+    @model_serializer(mode="wrap")
+    def serialize_model(self, handler):
+        optional_fields = set(["ttl"])
+        serialized = handler(self)
+        m = {}
+
+        for n, f in type(self).model_fields.items():
+            k = f.alias or n
+            val = serialized.get(k)
+
+            if val != UNSET_SENTINEL:
+                if val is not None or k not in optional_fields:
+                    m[k] = val
+
+        return m
+
+
+CreateAgentRequestLoadBalancerAgentsResponseType = Literal["weight_based",]
+
+
+class CreateAgentRequestLoadBalancerAgentsResponseModelsTypedDict(TypedDict):
+    model: str
+    r"""Model identifier for load balancing"""
+    weight: NotRequired[float]
+    r"""Weight assigned to this model for load balancing"""
+
+
+class CreateAgentRequestLoadBalancerAgentsResponseModels(BaseModel):
+    model: str
+    r"""Model identifier for load balancing"""
+
+    weight: Optional[float] = 0.5
+    r"""Weight assigned to this model for load balancing"""
+
+    @model_serializer(mode="wrap")
+    def serialize_model(self, handler):
+        optional_fields = set(["weight"])
+        serialized = handler(self)
+        m = {}
+
+        for n, f in type(self).model_fields.items():
+            k = f.alias or n
+            val = serialized.get(k)
+
+            if val != UNSET_SENTINEL:
+                if val is not None or k not in optional_fields:
+                    m[k] = val
+
+        return m
+
+
+class CreateAgentRequestLoadBalancerAgentsResponse1TypedDict(TypedDict):
+    type: CreateAgentRequestLoadBalancerAgentsResponseType
+    models: List[CreateAgentRequestLoadBalancerAgentsResponseModelsTypedDict]
+
+
+class CreateAgentRequestLoadBalancerAgentsResponse1(BaseModel):
+    type: CreateAgentRequestLoadBalancerAgentsResponseType
+
+    models: List[CreateAgentRequestLoadBalancerAgentsResponseModels]
+
+
+CreateAgentRequestFallbackModelConfigurationLoadBalancerTypedDict = (
+    CreateAgentRequestLoadBalancerAgentsResponse1TypedDict
+)
+r"""Load balancer configuration for the request."""
+
+
+CreateAgentRequestFallbackModelConfigurationLoadBalancer = (
+    CreateAgentRequestLoadBalancerAgentsResponse1
+)
+r"""Load balancer configuration for the request."""
+
+
+class CreateAgentRequestFallbackModelConfigurationTimeoutTypedDict(TypedDict):
+    r"""Timeout configuration to apply to the request. If the request exceeds the timeout, it will be retried or fallback to the next model if configured."""
+
+    call_timeout: float
+    r"""Timeout value in milliseconds"""
+
+
+class CreateAgentRequestFallbackModelConfigurationTimeout(BaseModel):
+    r"""Timeout configuration to apply to the request. If the request exceeds the timeout, it will be retried or fallback to the next model if configured."""
+
+    call_timeout: float
+    r"""Timeout value in milliseconds"""
+
+
 class CreateAgentRequestFallbackModelConfigurationParametersTypedDict(TypedDict):
     r"""Optional model parameters specific to this fallback model. Overrides primary model parameters if this fallback is used."""
 
@@ -2940,6 +4379,22 @@ class CreateAgentRequestFallbackModelConfigurationParametersTypedDict(TypedDict)
         List[CreateAgentRequestFallbackModelConfigurationGuardrailsTypedDict]
     ]
     r"""A list of guardrails to apply to the request."""
+    fallbacks: NotRequired[
+        List[CreateAgentRequestFallbackModelConfigurationFallbacksTypedDict]
+    ]
+    r"""Array of fallback models to use if primary model fails"""
+    retry: NotRequired[
+        CreateAgentRequestFallbackModelConfigurationAgentsResponseRetryTypedDict
+    ]
+    r"""Retry configuration for the request"""
+    cache: NotRequired[CreateAgentRequestFallbackModelConfigurationCacheTypedDict]
+    r"""Cache configuration for the request."""
+    load_balancer: NotRequired[
+        CreateAgentRequestFallbackModelConfigurationLoadBalancerTypedDict
+    ]
+    r"""Load balancer configuration for the request."""
+    timeout: NotRequired[CreateAgentRequestFallbackModelConfigurationTimeoutTypedDict]
+    r"""Timeout configuration to apply to the request. If the request exceeds the timeout, it will be retried or fallback to the next model if configured."""
 
 
 class CreateAgentRequestFallbackModelConfigurationParameters(BaseModel):
@@ -3031,77 +4486,102 @@ class CreateAgentRequestFallbackModelConfigurationParameters(BaseModel):
     ] = None
     r"""A list of guardrails to apply to the request."""
 
+    fallbacks: Optional[List[CreateAgentRequestFallbackModelConfigurationFallbacks]] = (
+        None
+    )
+    r"""Array of fallback models to use if primary model fails"""
+
+    retry: Optional[CreateAgentRequestFallbackModelConfigurationAgentsResponseRetry] = (
+        None
+    )
+    r"""Retry configuration for the request"""
+
+    cache: Optional[CreateAgentRequestFallbackModelConfigurationCache] = None
+    r"""Cache configuration for the request."""
+
+    load_balancer: Optional[
+        CreateAgentRequestFallbackModelConfigurationLoadBalancer
+    ] = None
+    r"""Load balancer configuration for the request."""
+
+    timeout: Optional[CreateAgentRequestFallbackModelConfigurationTimeout] = None
+    r"""Timeout configuration to apply to the request. If the request exceeds the timeout, it will be retried or fallback to the next model if configured."""
+
     @model_serializer(mode="wrap")
     def serialize_model(self, handler):
-        optional_fields = [
-            "audio",
-            "frequency_penalty",
-            "max_tokens",
-            "max_completion_tokens",
-            "logprobs",
-            "top_logprobs",
-            "n",
-            "presence_penalty",
-            "response_format",
-            "reasoning_effort",
-            "verbosity",
-            "seed",
-            "stop",
-            "stream_options",
-            "thinking",
-            "temperature",
-            "top_p",
-            "top_k",
-            "tool_choice",
-            "parallel_tool_calls",
-            "modalities",
-            "guardrails",
-        ]
-        nullable_fields = [
-            "audio",
-            "frequency_penalty",
-            "max_tokens",
-            "max_completion_tokens",
-            "logprobs",
-            "top_logprobs",
-            "n",
-            "presence_penalty",
-            "seed",
-            "stop",
-            "stream_options",
-            "temperature",
-            "top_p",
-            "top_k",
-            "modalities",
-        ]
-        null_default_fields = []
-
+        optional_fields = set(
+            [
+                "audio",
+                "frequency_penalty",
+                "max_tokens",
+                "max_completion_tokens",
+                "logprobs",
+                "top_logprobs",
+                "n",
+                "presence_penalty",
+                "response_format",
+                "reasoning_effort",
+                "verbosity",
+                "seed",
+                "stop",
+                "stream_options",
+                "thinking",
+                "temperature",
+                "top_p",
+                "top_k",
+                "tool_choice",
+                "parallel_tool_calls",
+                "modalities",
+                "guardrails",
+                "fallbacks",
+                "retry",
+                "cache",
+                "load_balancer",
+                "timeout",
+            ]
+        )
+        nullable_fields = set(
+            [
+                "audio",
+                "frequency_penalty",
+                "max_tokens",
+                "max_completion_tokens",
+                "logprobs",
+                "top_logprobs",
+                "n",
+                "presence_penalty",
+                "seed",
+                "stop",
+                "stream_options",
+                "temperature",
+                "top_p",
+                "top_k",
+                "modalities",
+            ]
+        )
         serialized = handler(self)
-
         m = {}
 
         for n, f in type(self).model_fields.items():
             k = f.alias or n
             val = serialized.get(k)
-            serialized.pop(k, None)
-
-            optional_nullable = k in optional_fields and k in nullable_fields
-            is_set = (
-                self.__pydantic_fields_set__.intersection({n})
-                or k in null_default_fields
-            )  # pylint: disable=no-member
-
-            if val is not None and val != UNSET_SENTINEL:
-                m[k] = val
-            elif val != UNSET_SENTINEL and (
-                not k in optional_fields or (optional_nullable and is_set)
-            ):
-                m[k] = val
+            is_nullable_and_explicitly_set = (
+                k in nullable_fields
+                and (self.__pydantic_fields_set__.intersection({n}))  # pylint: disable=no-member
+            )
+
+            if val != UNSET_SENTINEL:
+                if (
+                    val is not None
+                    or k not in optional_fields
+                    or is_nullable_and_explicitly_set
+                ):
+                    m[k] = val
 
         return m
 
 
-class CreateAgentRequestFallbackModelConfigurationRetryTypedDict(TypedDict):
+class CreateAgentRequestFallbackModelConfigurationAgentsRetryTypedDict(TypedDict):
     r"""Retry configuration for this fallback model. Allows customizing retry count (1-5) and HTTP status codes that trigger retries."""
 
     count: NotRequired[float]
@@ -3110,7 +4590,7 @@ class CreateAgentRequestFallbackModelConfigurationRetryTypedDict(TypedDict):
     r"""HTTP status codes that trigger retry logic"""
 
 
-class CreateAgentRequestFallbackModelConfigurationRetry(BaseModel):
+class CreateAgentRequestFallbackModelConfigurationAgentsRetry(BaseModel):
     r"""Retry configuration for this fallback model. Allows customizing retry count (1-5) and HTTP status codes that trigger retries."""
 
     count: Optional[float] = 3
@@ -3119,6 +4599,22 @@ class CreateAgentRequestFallbackModelConfigurationRetry(BaseModel):
     on_codes: Optional[List[float]] = None
     r"""HTTP status codes that trigger retry logic"""
 
+    @model_serializer(mode="wrap")
+    def serialize_model(self, handler):
+        optional_fields = set(["count", "on_codes"])
+        serialized = handler(self)
+        m = {}
+
+        for n, f in type(self).model_fields.items():
+            k = f.alias or n
+            val = serialized.get(k)
+
+            if val != UNSET_SENTINEL:
+                if val is not None or k not in optional_fields:
+                    m[k] = val
+
+        return m
+
 
 class CreateAgentRequestFallbackModelConfiguration2TypedDict(TypedDict):
     r"""Fallback model configuration with optional parameters and retry settings."""
@@ -3129,7 +4625,7 @@ class CreateAgentRequestFallbackModelConfiguration2TypedDict(TypedDict):
         CreateAgentRequestFallbackModelConfigurationParametersTypedDict
     ]
     r"""Optional model parameters specific to this fallback model. Overrides primary model parameters if this fallback is used."""
-    retry: NotRequired[CreateAgentRequestFallbackModelConfigurationRetryTypedDict]
+    retry: NotRequired[CreateAgentRequestFallbackModelConfigurationAgentsRetryTypedDict]
     r"""Retry configuration for this fallback model. Allows customizing retry count (1-5) and HTTP status codes that trigger retries."""
 
 
@@ -3142,9 +4638,25 @@ class CreateAgentRequestFallbackModelConfiguration2(BaseModel):
     parameters: Optional[CreateAgentRequestFallbackModelConfigurationParameters] = None
     r"""Optional model parameters specific to this fallback model. Overrides primary model parameters if this fallback is used."""
 
-    retry: Optional[CreateAgentRequestFallbackModelConfigurationRetry] = None
+    retry: Optional[CreateAgentRequestFallbackModelConfigurationAgentsRetry] = None
     r"""Retry configuration for this fallback model. Allows customizing retry count (1-5) and HTTP status codes that trigger retries."""
 
+    @model_serializer(mode="wrap")
+    def serialize_model(self, handler):
+        optional_fields = set(["parameters", "retry"])
+        serialized = handler(self)
+        m = {}
+
+        for n, f in type(self).model_fields.items():
+            k = f.alias or n
+            val = serialized.get(k)
+
+            if val != UNSET_SENTINEL:
+                if val is not None or k not in optional_fields:
+                    m[k] = val
+
+        return m
+
 
 CreateAgentRequestFallbackModelConfigurationTypedDict = TypeAliasType(
     "CreateAgentRequestFallbackModelConfigurationTypedDict",
@@ -3195,31 +4707,28 @@ class Model(BaseModel):
 
     @model_serializer(mode="wrap")
     def serialize_model(self, handler):
-        optional_fields = ["integration_id", "parameters", "retry", "fallback_models"]
-        nullable_fields = ["integration_id", "fallback_models"]
-        null_default_fields = []
-
+        optional_fields = set(
+            ["integration_id", "parameters", "retry", "fallback_models"]
+        )
+        nullable_fields = set(["integration_id", "fallback_models"])
         serialized = handler(self)
-
         m = {}
 
         for n, f in type(self).model_fields.items():
             k = f.alias or n
             val = serialized.get(k)
-            serialized.pop(k, None)
-
-            optional_nullable = k in optional_fields and k in nullable_fields
-            is_set = (
-                self.__pydantic_fields_set__.intersection({n})
-                or k in null_default_fields
-            )  # pylint: disable=no-member
-
-            if val is not None and val != UNSET_SENTINEL:
-                m[k] = val
-            elif val != UNSET_SENTINEL and (
-                not k in optional_fields or (optional_nullable and is_set)
-            ):
-                m[k] = val
+            is_nullable_and_explicitly_set = (
+                k in nullable_fields
+                and (self.__pydantic_fields_set__.intersection({n}))  # pylint: disable=no-member
+            )
+
+            if val != UNSET_SENTINEL:
+                if (
+                    val is not None
+                    or k not in optional_fields
+                    or is_nullable_and_explicitly_set
+                ):
+                    m[k] = val
 
         return m
 
@@ -3238,6 +4747,22 @@ class CreateAgentRequestTeamOfAgents(BaseModel):
     role: Optional[str] = None
     r"""The role of the agent in this context. This is used to give extra information to the leader to help it decide which agent to hand off to."""
 
+    @model_serializer(mode="wrap")
+    def serialize_model(self, handler):
+        optional_fields = set(["role"])
+        serialized = handler(self)
+        m = {}
+
+        for n, f in type(self).model_fields.items():
+            k = f.alias or n
+            val = serialized.get(k)
+
+            if val != UNSET_SENTINEL:
+                if val is not None or k not in optional_fields:
+                    m[k] = val
+
+        return m
+
 
 class CreateAgentRequestMetricsTypedDict(TypedDict):
     total_cost: NotRequired[float]
@@ -3246,6 +4771,22 @@ class CreateAgentRequestMetricsTypedDict(TypedDict):
 class CreateAgentRequestMetrics(BaseModel):
     total_cost: Optional[float] = 0
 
+    @model_serializer(mode="wrap")
+    def serialize_model(self, handler):
+        optional_fields = set(["total_cost"])
+        serialized = handler(self)
+        m = {}
+
+        for n, f in type(self).model_fields.items():
+            k = f.alias or n
+            val = serialized.get(k)
+
+            if val != UNSET_SENTINEL:
+                if val is not None or k not in optional_fields:
+                    m[k] = val
+
+        return m
+
 
 class CreateAgentRequestKnowledgeBasesTypedDict(TypedDict):
     knowledge_id: str
@@ -3367,42 +4908,39 @@ class CreateAgentRequestResponseBody(BaseModel):
 
     @model_serializer(mode="wrap")
     def serialize_model(self, handler):
-        optional_fields = [
-            "created_by_id",
-            "updated_by_id",
-            "created",
-            "updated",
-            "system_prompt",
-            "settings",
-            "version_hash",
-            "metrics",
-            "variables",
-            "knowledge_bases",
-            "source",
-        ]
-        nullable_fields = ["created_by_id", "updated_by_id"]
-        null_default_fields = []
-
+        optional_fields = set(
+            [
+                "created_by_id",
+                "updated_by_id",
+                "created",
+                "updated",
+                "system_prompt",
+                "settings",
+                "version_hash",
+                "metrics",
+                "variables",
+                "knowledge_bases",
+                "source",
+            ]
+        )
+        nullable_fields = set(["created_by_id", "updated_by_id"])
         serialized = handler(self)
-
         m = {}
 
         for n, f in type(self).model_fields.items():
             k = f.alias or n
             val = serialized.get(k)
-            serialized.pop(k, None)
-
-            optional_nullable = k in optional_fields and k in nullable_fields
-            is_set = (
-                self.__pydantic_fields_set__.intersection({n})
-                or k in null_default_fields
-            )  # pylint: disable=no-member
-
-            if val is not None and val != UNSET_SENTINEL:
-                m[k] = val
-            elif val != UNSET_SENTINEL and (
-                not k in optional_fields or (optional_nullable and is_set)
-            ):
-                m[k] = val
+            is_nullable_and_explicitly_set = (
+                k in nullable_fields
+                and (self.__pydantic_fields_set__.intersection({n}))  # pylint: disable=no-member
+            )
+
+            if val != UNSET_SENTINEL:
+                if (
+                    val is not None
+                    or k not in optional_fields
+                    or is_nullable_and_explicitly_set
+                ):
+                    m[k] = val
 
         return m