unique-toolkit 2026.42.0.dev0__tar.gz → 2026.42.0.dev2__tar.gz

{unique_toolkit-2026.42.0.dev0 → unique_toolkit-2026.42.0.dev2}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.3
 Name: unique-toolkit
-Version: 2026.42.0.dev0
+Version: 2026.42.0.dev2
 Summary: 
 Author: Cedric Klinkert, Martin Fadler, Sadique Sheik, Fabian Schläpfer, Pascal Hauri, Ejiroghene Oki-Peter, Rami Azouz, Ricardo Visini, Sanzio Monti
 Author-email: Cedric Klinkert <cedric.klinkert@unique.ch>, Martin Fadler <martin.fadler@unique.ch>, Sadique Sheik <sadique@unique.ch>, Fabian Schläpfer <fabian@unique.ch>, Pascal Hauri <pascal@unique.ch>, Ejiroghene Oki-Peter <oki-peter.ejiroghene@unique.ai>, Rami Azouz <rami.ext@unique.ch>, Ricardo Visini <ricardo.visini@unique.ch>, Sanzio Monti <sanzio.monti@unique.ai>
@@ -16,7 +16,7 @@ Requires-Dist: sseclient>=0.0.27,<0.0.28
 Requires-Dist: openai>=1.105.0,<3
 Requires-Dist: platformdirs>=4.0.0,<5
 Requires-Dist: pillow>=12.1.1
-Requires-Dist: unique-sdk>=2026.42.0.dev0,<2026.42.0rc0
+Requires-Dist: unique-sdk>=2026.42.0.dev1,<2026.42.0rc0
 Requires-Dist: jambo>=0.1.2,<0.2
 Requires-Dist: jinja2>=3.1.6,<4
 Requires-Dist: docxtpl>=0.20.1,<0.21

{unique_toolkit-2026.42.0.dev0 → unique_toolkit-2026.42.0.dev2}/pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "unique_toolkit"
-version = "2026.42.0.dev0"
+version = "2026.42.0.dev2"
 description = ""
 readme = "README.md"
 requires-python = ">=3.12"
@@ -28,7 +28,7 @@ dependencies = [
     "openai>=1.105.0,<3",
     "platformdirs>=4.0.0,<5",
     "pillow>=12.1.1",
-    "unique-sdk>=2026.42.0.dev0,<2026.42.0rc0",
+    "unique-sdk>=2026.42.0.dev1,<2026.42.0rc0",
     "jambo>=0.1.2,<0.2",
     "jinja2>=3.1.6,<4",
     "docxtpl>=0.20.1,<0.21",

{unique_toolkit-2026.42.0.dev0 → unique_toolkit-2026.42.0.dev2}/unique_toolkit/language_model/infos.py

@@ -68,6 +68,7 @@ class LanguageModelName(StrEnum):
     AZURE_GPT_6_ASTRA_2026_0903 = "AZURE_GPT_6_ASTRA_2026_0903"
     AZURE_GPT_6_LUNA_2026_0922 = "AZURE_GPT_6_LUNA_2026_0922"
     AZURE_GPT_6_SOL_2026_0922 = "AZURE_GPT_6_SOL_2026_0922"
+    AZURE_GPT_61_SOL_2026_0929 = "AZURE_GPT_61_SOL_2026_0929"
     AZURE_GPT_4o_2024_0513 = "AZURE_GPT_4o_2024_0513"
     AZURE_GPT_4o_2024_0806 = "AZURE_GPT_4o_2024_0806"
     AZURE_GPT_4o_2024_1120 = "AZURE_GPT_4o_2024_1120"
@@ -91,6 +92,7 @@ class LanguageModelName(StrEnum):
     ANTHROPIC_CLAUDE_SONNET_4_5 = "litellm:anthropic-claude-sonnet-4-5"
     ANTHROPIC_CLAUDE_SONNET_4_6 = "litellm:anthropic-claude-sonnet-4-6"
     ANTHROPIC_CLAUDE_SONNET_5 = "litellm:anthropic-claude-sonnet-5"
+    ANTHROPIC_CLAUDE_SONNET_5_5 = "litellm:anthropic-claude-sonnet-5-5"
     ANTHROPIC_CLAUDE_OPUS_4 = "litellm:anthropic-claude-opus-4"
     ANTHROPIC_CLAUDE_OPUS_4_1 = "litellm:anthropic-claude-opus-4-1"
     ANTHROPIC_CLAUDE_OPUS_4_5 = "litellm:anthropic-claude-opus-4-5"
@@ -137,6 +139,7 @@ class LanguageModelName(StrEnum):
     LITELLM_OPENAI_GPT_6_ASTRA = "litellm:openai-gpt-6-astra"
     LITELLM_OPENAI_GPT_6_LUNA = "litellm:openai-gpt-6-luna"
     LITELLM_OPENAI_GPT_6_SOL = "litellm:openai-gpt-6-sol"
+    LITELLM_OPENAI_GPT_61_SOL = "litellm:openai-gpt-6.1-sol"
     LITELLM_OPENAI_O1 = "litellm:openai-o1"
     LITELLM_OPENAI_O3 = "litellm:openai-o3"
     LITELLM_OPENAI_O3_DEEP_RESEARCH = "litellm:openai-o3-deep-research"
@@ -160,6 +163,7 @@ class LanguageModelName(StrEnum):
     LITELLM_QWEN_3_THINKING = "litellm:qwen-3-235B-A22B-thinking"
     VERTEX_CLAUDE_SONNET_4_6 = "litellm:vertex-claude-sonnet-4-6"
     VERTEX_CLAUDE_SONNET_5 = "litellm:vertex-claude-sonnet-5"
+    VERTEX_CLAUDE_SONNET_5_5 = "litellm:vertex-claude-sonnet-5-5"
     VERTEX_CLAUDE_OPUS_4_6 = "litellm:vertex-claude-opus-4-6"
     VERTEX_CLAUDE_OPUS_4_7 = "litellm:vertex-claude-opus-4-7"
     VERTEX_CLAUDE_OPUS_4_8 = "litellm:vertex-claude-opus-4-8"
@@ -235,6 +239,7 @@ def get_encoder_name(model_name: LanguageModelName) -> EncoderName:
             | LMN.AZURE_GPT_6_ASTRA_2026_0903
             | LMN.AZURE_GPT_6_LUNA_2026_0922
             | LMN.AZURE_GPT_6_SOL_2026_0922
+            | LMN.AZURE_GPT_61_SOL_2026_0929
             | LMN.AZURE_MODEL_ROUTER_2025_1118
             | LMN.LITELLM_OPENAI_GPT_5
             | LMN.LITELLM_OPENAI_GPT_5_MINI
@@ -255,6 +260,7 @@ def get_encoder_name(model_name: LanguageModelName) -> EncoderName:
             | LMN.LITELLM_OPENAI_GPT_6_ASTRA
             | LMN.LITELLM_OPENAI_GPT_6_LUNA
             | LMN.LITELLM_OPENAI_GPT_6_SOL
+            | LMN.LITELLM_OPENAI_GPT_61_SOL
             | LMN.LITELLM_OPENAI_O1
             | LMN.LITELLM_OPENAI_O3
             | LMN.LITELLM_OPENAI_O3_DEEP_RESEARCH
@@ -1846,6 +1852,33 @@ class LanguageModelInfo(BaseModel):
                     ),
                     supported_reasoning_efforts=[],
                 )
+            case (
+                LanguageModelName.ANTHROPIC_CLAUDE_SONNET_5_5
+                | LanguageModelName.VERTEX_CLAUDE_SONNET_5_5
+            ):
+                return cls(
+                    name=model_name,
+                    capabilities=[
+                        ModelCapabilities.FUNCTION_CALLING,
+                        ModelCapabilities.STREAMING,
+                        ModelCapabilities.VISION,
+                        ModelCapabilities.REASONING,
+                    ],
+                    provider=LanguageModelProvider.LITELLM,
+                    family=ModelFamily.ANTHROPIC,
+                    version="claude-sonnet-5-5",
+                    encoder_name=EncoderName.O200K_BASE,  # TODO: Update encoder with litellm
+                    token_limits=LanguageModelTokenLimits(
+                        # TODO: Remove the 1.3 adjustment once a proper Claude tokenizer is implemented. UN-24123
+                        # 1M context / 128K max output. https://platform.claude.com/docs/en/models/sonnet-5-5/overview
+                        token_limit_input=int(1_000_000 / 1.3),
+                        token_limit_output=int(128_000 / 1.3),
+                    ),
+                    info_cutoff_at=date(2026, 6, 1),
+                    published_at=date(2026, 9, 28),
+                    supported_reasoning_efforts=[],
+                    supports_forced_tool_choice=False,
+                )
             case LanguageModelName.ANTHROPIC_CLAUDE_OPUS_4:
                 return cls(
                     name=model_name,
@@ -2984,6 +3017,55 @@ class LanguageModelInfo(BaseModel):
                         "xhigh",
                     ],
                 )
+            case (
+                LanguageModelName.AZURE_GPT_61_SOL_2026_0929
+                | LanguageModelName.LITELLM_OPENAI_GPT_61_SOL
+            ):
+                # https://developers.openai.com/api/docs/models/gpt-6.1-sol
+                # Foundry GA 2026-09-29:
+                # https://techcommunity.microsoft.com/blog/azure-ai-foundry-blog/introducing-gpt-6-1-sol-in-microsoft-foundry-advanced-intelligence-optimized-for/4560811
+                # Context 1,050,000 with 128,000 max output → 922,000 max input.
+                # `none` is not a reasoning effort. `max` is documented.
+                provider = (
+                    LanguageModelProvider.AZURE
+                    if model_name is LanguageModelName.AZURE_GPT_61_SOL_2026_0929
+                    else LanguageModelProvider.LITELLM
+                )
+                return cls(
+                    name=model_name,
+                    provider=provider,
+                    family=ModelFamily.OPENAI,
+                    version="2026-09-29",
+                    encoder_name=EncoderName.O200K_BASE,
+                    capabilities=[
+                        ModelCapabilities.CHAT_COMPLETIONS_API,
+                        ModelCapabilities.FUNCTION_CALLING,
+                        ModelCapabilities.PARALLEL_FUNCTION_CALLING,
+                        ModelCapabilities.REASONING,
+                        ModelCapabilities.RESPONSES_API,
+                        ModelCapabilities.STREAMING,
+                        ModelCapabilities.STRUCTURED_OUTPUT,
+                        ModelCapabilities.VISION,
+                    ],
+                    token_limits=LanguageModelTokenLimits(
+                        token_limit_input=922_000, token_limit_output=128_000
+                    ),
+                    info_cutoff_at=date(2026, 4, 30),
+                    published_at=date(2026, 9, 29),
+                    temperature_bounds=TemperatureBounds(
+                        min_temperature=0.0, max_temperature=1.0
+                    ),
+                    default_options={
+                        "reasoning_effort": "medium",
+                    },
+                    supported_reasoning_efforts=[
+                        "low",
+                        "medium",
+                        "high",
+                        "xhigh",
+                        "max",
+                    ],
+                )
             case LanguageModelName.LITELLM_OPENAI_O1:
                 return cls(
                     name=model_name,

{unique_toolkit-2026.42.0.dev0 → unique_toolkit-2026.42.0.dev2}/unique_toolkit/language_model/schemas.py

@@ -793,9 +793,9 @@ class LanguageModelResponse(BaseModel):
 
 
 # The OpenAI SDK's ReasoningEffort type alias is generated from an older OpenAPI spec and is
-# missing values that the API actually supports (e.g. "xhigh", "none"). We define our own
+# missing values that the API actually supports (e.g. "xhigh", "none", "max"). We define our own
 # complete type here as the source of truth.
-ReasoningEffort = Literal["none", "minimal", "low", "medium", "high", "xhigh"]
+ReasoningEffort = Literal["none", "minimal", "low", "medium", "high", "xhigh", "max"]
 
 # Ordered from lowest to highest reasoning effort. Used wherever effort levels must be
 # compared or ranked (e.g. picking the maximum across multiple skill hints).
@@ -806,18 +806,19 @@ REASONING_EFFORT_ORDER: tuple[ReasoningEffort, ...] = (
     "medium",
     "high",
     "xhigh",
+    "max",
 )
 
 
 def to_reasoning_effort(value: str) -> ReasoningEffort:
     """Narrow a raw string to ReasoningEffort, raising ValueError for unrecognised values."""
     match value:
-        case "none" | "minimal" | "low" | "medium" | "high" | "xhigh":
+        case "none" | "minimal" | "low" | "medium" | "high" | "xhigh" | "max":
             return value
         case _:
             raise ValueError(
                 f"Unknown reasoning_effort {value!r}. "
-                f"Supported values: none, minimal, low, medium, high, xhigh."
+                f"Supported values: none, minimal, low, medium, high, xhigh, max."
             )