kiln-ai 0.15.0__py3-none-any.whl → 0.17.0__py3-none-any.whl

kiln_ai/adapters/ml_model_list.py

@@ -3,7 +3,11 @@ from typing import Dict, List, Literal
 
 from pydantic import BaseModel
 
-from kiln_ai.datamodel import StructuredOutputMode
+from kiln_ai.datamodel.datamodel_enums import (
+    ChatStrategy,
+    ModelProviderName,
+    StructuredOutputMode,
+)
 
 """
 Provides model configuration and management for various LLM providers and models.
@@ -12,28 +16,6 @@ including configuration, validation, and instantiation of language models.
 """
 
 
-class ModelProviderName(str, Enum):
-    """
-    Enumeration of supported AI model providers.
-    """
-
-    openai = "openai"
-    groq = "groq"
-    amazon_bedrock = "amazon_bedrock"
-    ollama = "ollama"
-    openrouter = "openrouter"
-    fireworks_ai = "fireworks_ai"
-    kiln_fine_tune = "kiln_fine_tune"
-    kiln_custom_registry = "kiln_custom_registry"
-    openai_compatible = "openai_compatible"
-    anthropic = "anthropic"
-    gemini_api = "gemini_api"
-    azure_openai = "azure_openai"
-    huggingface = "huggingface"
-    vertex = "vertex"
-    together_ai = "together_ai"
-
-
 class ModelFamily(str, Enum):
     """
     Enumeration of supported model families/architectures.
@@ -102,6 +84,8 @@ class ModelName(str, Enum):
     claude_3_5_sonnet = "claude_3_5_sonnet"
     claude_3_7_sonnet = "claude_3_7_sonnet"
     claude_3_7_sonnet_thinking = "claude_3_7_sonnet_thinking"
+    claude_sonnet_4 = "claude_sonnet_4"
+    claude_opus_4 = "claude_opus_4"
     gemini_1_5_flash = "gemini_1_5_flash"
     gemini_1_5_flash_8b = "gemini_1_5_flash_8b"
     gemini_1_5_pro = "gemini_1_5_pro"
@@ -126,6 +110,22 @@ class ModelName(str, Enum):
     deepseek_r1_distill_llama_8b = "deepseek_r1_distill_llama_8b"
     dolphin_2_9_8x22b = "dolphin_2_9_8x22b"
     grok_2 = "grok_2"
+    qwen_3_0p6b = "qwen_3_0p6b"
+    qwen_3_0p6b_no_thinking = "qwen_3_0p6b_no_thinking"
+    qwen_3_1p7b = "qwen_3_1p7b"
+    qwen_3_1p7b_no_thinking = "qwen_3_1p7b_no_thinking"
+    qwen_3_4b = "qwen_3_4b"
+    qwen_3_4b_no_thinking = "qwen_3_4b_no_thinking"
+    qwen_3_8b = "qwen_3_8b"
+    qwen_3_8b_no_thinking = "qwen_3_8b_no_thinking"
+    qwen_3_14b = "qwen_3_14b"
+    qwen_3_14b_no_thinking = "qwen_3_14b_no_thinking"
+    qwen_3_30b_a3b = "qwen_3_30b_a3b"
+    qwen_3_30b_a3b_no_thinking = "qwen_3_30b_a3b_no_thinking"
+    qwen_3_32b = "qwen_3_32b"
+    qwen_3_32b_no_thinking = "qwen_3_32b_no_thinking"
+    qwen_3_235b_a22b = "qwen_3_235b_a22b"
+    qwen_3_235b_a22b_no_thinking = "qwen_3_235b_a22b_no_thinking"
 
 
 class ModelParserID(str, Enum):
@@ -134,6 +134,15 @@ class ModelParserID(str, Enum):
     """
 
     r1_thinking = "r1_thinking"
+    optional_r1_thinking = "optional_r1_thinking"
+
+
+class ModelFormatterID(str, Enum):
+    """
+    Enumeration of supported model formatters.
+    """
+
+    qwen3_style_no_think = "qwen3_style_no_think"
 
 
 class KilnModelProvider(BaseModel):
@@ -149,18 +158,23 @@ class KilnModelProvider(BaseModel):
         structured_output_mode: The mode we should use to call the model for structured output, if it was trained with structured output.
         parser: A parser to use for the model, if applicable
         reasoning_capable: Whether the model is designed to output thinking in a structured format (eg <think></think>). If so we don't use COT across 2 calls, and ask for thinking and final response in the same call.
+        tuned_chat_strategy: Used when a model is finetuned with a specific chat strategy, and it's best to use it at call time.
     """
 
     name: ModelProviderName
     model_id: str | None = None
     supports_structured_output: bool = True
     supports_data_gen: bool = True
+    suggested_for_data_gen: bool = False
     untested_model: bool = False
     provider_finetune_id: str | None = None
     structured_output_mode: StructuredOutputMode = StructuredOutputMode.default
     parser: ModelParserID | None = None
+    formatter: ModelFormatterID | None = None
     reasoning_capable: bool = False
     supports_logprobs: bool = False
+    suggested_for_evals: bool = False
+    tuned_chat_strategy: ChatStrategy | None = None
 
     # TODO P1: Need a more generalized way to handle custom provider parameters.
     # Making them quite declarative here for now, isolating provider specific logic
@@ -205,16 +219,22 @@ built_in_models: List[KilnModel] = [
                 provider_finetune_id="gpt-4.1-2025-04-14",
                 structured_output_mode=StructuredOutputMode.json_schema,
                 supports_logprobs=True,
+                suggested_for_data_gen=True,
+                suggested_for_evals=True,
             ),
             KilnModelProvider(
                 name=ModelProviderName.openrouter,
                 model_id="openai/gpt-4.1",
                 structured_output_mode=StructuredOutputMode.json_schema,
                 supports_logprobs=True,
+                suggested_for_data_gen=True,
+                suggested_for_evals=True,
             ),
             KilnModelProvider(
                 name=ModelProviderName.azure_openai,
                 model_id="gpt-4.1",
+                suggested_for_data_gen=True,
+                suggested_for_evals=True,
             ),
         ],
     ),
@@ -230,16 +250,22 @@ built_in_models: List[KilnModel] = [
                 provider_finetune_id="gpt-4.1-mini-2025-04-14",
                 structured_output_mode=StructuredOutputMode.json_schema,
                 supports_logprobs=True,
+                suggested_for_data_gen=True,
+                suggested_for_evals=True,
             ),
             KilnModelProvider(
                 name=ModelProviderName.openrouter,
                 model_id="openai/gpt-4.1-mini",
                 structured_output_mode=StructuredOutputMode.json_schema,
                 supports_logprobs=True,
+                suggested_for_data_gen=True,
+                suggested_for_evals=True,
             ),
             KilnModelProvider(
                 name=ModelProviderName.azure_openai,
                 model_id="gpt-4.1-mini",
+                suggested_for_data_gen=True,
+                suggested_for_evals=True,
             ),
         ],
     ),
@@ -625,11 +651,13 @@ built_in_models: List[KilnModel] = [
                 name=ModelProviderName.openrouter,
                 structured_output_mode=StructuredOutputMode.function_calling,
                 model_id="anthropic/claude-3.7-sonnet",
+                suggested_for_data_gen=True,
             ),
             KilnModelProvider(
                 name=ModelProviderName.anthropic,
                 model_id="claude-3-7-sonnet-20250219",
                 structured_output_mode=StructuredOutputMode.function_calling,
+                suggested_for_data_gen=True,
             ),
         ],
     ),
@@ -656,6 +684,46 @@ built_in_models: List[KilnModel] = [
             ),
         ],
     ),
+    # Claude Sonnet 4
+    KilnModel(
+        family=ModelFamily.claude,
+        name=ModelName.claude_sonnet_4,
+        friendly_name="Claude Sonnet 4",
+        providers=[
+            KilnModelProvider(
+                name=ModelProviderName.openrouter,
+                model_id="anthropic/claude-sonnet-4",
+                structured_output_mode=StructuredOutputMode.function_calling,
+                suggested_for_data_gen=True,
+                suggested_for_evals=True,
+            ),
+            KilnModelProvider(
+                name=ModelProviderName.anthropic,
+                model_id="claude-sonnet-4-20250514",
+                structured_output_mode=StructuredOutputMode.function_calling,
+                suggested_for_data_gen=True,
+                suggested_for_evals=True,
+            ),
+        ],
+    ),
+    # Claude Opus 4
+    KilnModel(
+        family=ModelFamily.claude,
+        name=ModelName.claude_opus_4,
+        friendly_name="Claude Opus 4",
+        providers=[
+            KilnModelProvider(
+                name=ModelProviderName.openrouter,
+                model_id="anthropic/claude-opus-4",
+                structured_output_mode=StructuredOutputMode.function_calling,
+            ),
+            KilnModelProvider(
+                name=ModelProviderName.anthropic,
+                model_id="claude-opus-4-20250514",
+                structured_output_mode=StructuredOutputMode.function_calling,
+            ),
+        ],
+    ),
     # Gemini 2.5 Pro
     KilnModel(
         family=ModelFamily.gemini,
@@ -666,16 +734,22 @@ built_in_models: List[KilnModel] = [
                 name=ModelProviderName.openrouter,
                 model_id="google/gemini-2.5-pro-preview-03-25",
                 structured_output_mode=StructuredOutputMode.json_schema,
+                suggested_for_data_gen=True,
+                suggested_for_evals=True,
             ),
             KilnModelProvider(
                 name=ModelProviderName.gemini_api,
                 model_id="gemini-2.5-pro-preview-03-25",
                 structured_output_mode=StructuredOutputMode.json_schema,
+                suggested_for_data_gen=True,
+                suggested_for_evals=True,
             ),
             KilnModelProvider(
                 name=ModelProviderName.vertex,
                 model_id="gemini-2.5-pro-preview-03-25",
                 structured_output_mode=StructuredOutputMode.json_schema,
+                suggested_for_data_gen=True,
+                suggested_for_evals=True,
             ),
         ],
     ),
@@ -870,7 +944,7 @@ built_in_models: List[KilnModel] = [
                 model_id="meta-llama/Meta-Llama-3.1-8B-Instruct-Turbo",
                 supports_data_gen=False,
                 structured_output_mode=StructuredOutputMode.function_calling_weak,
-                provider_finetune_id="meta-llama/Meta-Llama-3.1-8B-Instruct",
+                provider_finetune_id="meta-llama/Meta-Llama-3.1-8B-Instruct-Reference",
             ),
         ],
     ),
@@ -889,8 +963,7 @@ built_in_models: List[KilnModel] = [
             KilnModelProvider(
                 name=ModelProviderName.openrouter,
                 supports_data_gen=False,
-                # Need to not pass "strict=True" to the function call to get this to work with logprobs for some reason. Openrouter issue.
-                structured_output_mode=StructuredOutputMode.function_calling_weak,
+                structured_output_mode=StructuredOutputMode.json_schema,
                 model_id="meta-llama/llama-3.1-70b-instruct",
                 supports_logprobs=True,
                 logprobs_openrouter_options=True,
@@ -911,7 +984,7 @@ built_in_models: List[KilnModel] = [
                 model_id="meta-llama/Meta-Llama-3.1-70B-Instruct-Turbo",
                 supports_data_gen=False,
                 structured_output_mode=StructuredOutputMode.function_calling_weak,
-                provider_finetune_id="meta-llama/Meta-Llama-3.1-70B-Instruct",
+                provider_finetune_id="meta-llama/Meta-Llama-3.1-70B-Instruct-Reference",
             ),
         ],
     ),
@@ -1018,10 +1091,6 @@ built_in_models: List[KilnModel] = [
                 supports_structured_output=False,
                 supports_data_gen=False,
             ),
-            KilnModelProvider(
-                name=ModelProviderName.together_ai,
-                provider_finetune_id="meta-llama/Llama-3.2-1B-Instruct",
-            ),
         ],
     ),
     # Llama 3.2 3B
@@ -1058,7 +1127,6 @@ built_in_models: List[KilnModel] = [
                 model_id="meta-llama/Llama-3.2-3B-Instruct-Turbo",
                 supports_structured_output=False,
                 supports_data_gen=False,
-                provider_finetune_id="meta-llama/Llama-3.2-3B-Instruct",
             ),
         ],
     ),
@@ -1068,10 +1136,6 @@ built_in_models: List[KilnModel] = [
         name=ModelName.llama_3_2_11b,
         friendly_name="Llama 3.2 11B",
         providers=[
-            KilnModelProvider(
-                name=ModelProviderName.groq,
-                model_id="llama-3.2-11b-vision-preview",
-            ),
             KilnModelProvider(
                 name=ModelProviderName.openrouter,
                 # Best mode, but fails to often to enable without warning
@@ -1112,10 +1176,6 @@ built_in_models: List[KilnModel] = [
         name=ModelName.llama_3_2_90b,
         friendly_name="Llama 3.2 90B",
         providers=[
-            KilnModelProvider(
-                name=ModelProviderName.groq,
-                model_id="llama-3.2-90b-vision-preview",
-            ),
             KilnModelProvider(
                 name=ModelProviderName.openrouter,
                 structured_output_mode=StructuredOutputMode.json_instruction_and_object,
@@ -1388,11 +1448,6 @@ built_in_models: List[KilnModel] = [
                 structured_output_mode=StructuredOutputMode.json_instruction_and_object,
                 model_id="google/gemma-3-27b-it",
             ),
-            KilnModelProvider(
-                name=ModelProviderName.huggingface,
-                model_id="google/gemma-3-27b-it",
-                structured_output_mode=StructuredOutputMode.json_instructions,
-            ),
         ],
     ),
     # Mixtral 8x7B
@@ -1578,6 +1633,8 @@ built_in_models: List[KilnModel] = [
                 reasoning_capable=True,
                 r1_openrouter_options=True,
                 require_openrouter_reasoning=True,
+                suggested_for_data_gen=True,
+                suggested_for_evals=True,
             ),
             KilnModelProvider(
                 name=ModelProviderName.fireworks_ai,
@@ -1585,6 +1642,8 @@ built_in_models: List[KilnModel] = [
                 parser=ModelParserID.r1_thinking,
                 structured_output_mode=StructuredOutputMode.json_instructions,
                 reasoning_capable=True,
+                suggested_for_data_gen=True,
+                suggested_for_evals=True,
             ),
             KilnModelProvider(
                 # I want your RAM
@@ -1593,6 +1652,8 @@ built_in_models: List[KilnModel] = [
                 parser=ModelParserID.r1_thinking,
                 structured_output_mode=StructuredOutputMode.json_instructions,
                 reasoning_capable=True,
+                suggested_for_data_gen=True,
+                suggested_for_evals=True,
             ),
             KilnModelProvider(
                 name=ModelProviderName.together_ai,
@@ -1600,6 +1661,8 @@ built_in_models: List[KilnModel] = [
                 structured_output_mode=StructuredOutputMode.json_instructions,
                 parser=ModelParserID.r1_thinking,
                 reasoning_capable=True,
+                suggested_for_data_gen=True,
+                suggested_for_evals=True,
             ),
         ],
     ),
@@ -1819,4 +1882,452 @@ built_in_models: List[KilnModel] = [
             ),
         ],
     ),
+    # Qwen 3 0.6B
+    KilnModel(
+        family=ModelFamily.qwen,
+        name=ModelName.qwen_3_0p6b,
+        friendly_name="Qwen 3 0.6B",
+        providers=[
+            KilnModelProvider(
+                name=ModelProviderName.openrouter,
+                model_id="qwen/qwen3-0.6b-04-28:free",
+                structured_output_mode=StructuredOutputMode.json_instructions,
+                reasoning_capable=True,
+                require_openrouter_reasoning=True,
+                r1_openrouter_options=True,
+                parser=ModelParserID.r1_thinking,
+                supports_data_gen=False,
+            ),
+            KilnModelProvider(
+                name=ModelProviderName.ollama,
+                model_id="qwen3:0.6b",
+                supports_data_gen=False,
+                reasoning_capable=True,
+                structured_output_mode=StructuredOutputMode.json_schema,
+            ),
+        ],
+    ),
+    # Qwen 3 0.6B Non-Thinking -- not respecting /no_think tag, skipping
+    # Qwen 3 1.7B
+    KilnModel(
+        family=ModelFamily.qwen,
+        name=ModelName.qwen_3_1p7b,
+        friendly_name="Qwen 3 1.7B",
+        providers=[
+            KilnModelProvider(
+                name=ModelProviderName.openrouter,
+                model_id="qwen/qwen3-1.7b:free",
+                structured_output_mode=StructuredOutputMode.json_instructions,
+                reasoning_capable=True,
+                require_openrouter_reasoning=True,
+                r1_openrouter_options=True,
+                parser=ModelParserID.r1_thinking,
+                supports_data_gen=False,
+            ),
+            KilnModelProvider(
+                name=ModelProviderName.ollama,
+                model_id="qwen3:1.7b",
+                supports_data_gen=False,
+                reasoning_capable=True,
+                structured_output_mode=StructuredOutputMode.json_schema,
+            ),
+        ],
+    ),
+    # Qwen 3 1.7B Non-Thinking
+    KilnModel(
+        family=ModelFamily.qwen,
+        name=ModelName.qwen_3_1p7b_no_thinking,
+        friendly_name="Qwen 3 1.7B Non-Thinking",
+        providers=[
+            KilnModelProvider(
+                name=ModelProviderName.openrouter,
+                model_id="qwen/qwen3-1.7b:free",
+                structured_output_mode=StructuredOutputMode.json_instructions,
+                formatter=ModelFormatterID.qwen3_style_no_think,
+                supports_data_gen=False,
+                parser=ModelParserID.optional_r1_thinking,
+            ),
+            KilnModelProvider(
+                name=ModelProviderName.ollama,
+                model_id="qwen3:1.7b",
+                formatter=ModelFormatterID.qwen3_style_no_think,
+                supports_data_gen=False,
+                structured_output_mode=StructuredOutputMode.json_schema,
+            ),
+        ],
+    ),
+    # Qwen 3 4B
+    KilnModel(
+        family=ModelFamily.qwen,
+        name=ModelName.qwen_3_4b,
+        friendly_name="Qwen 3 4B",
+        providers=[
+            KilnModelProvider(
+                name=ModelProviderName.openrouter,
+                model_id="qwen/qwen3-4b:free",
+                structured_output_mode=StructuredOutputMode.json_instructions,
+                reasoning_capable=True,
+                require_openrouter_reasoning=True,
+                r1_openrouter_options=True,
+                parser=ModelParserID.r1_thinking,
+                supports_data_gen=False,
+            ),
+            KilnModelProvider(
+                name=ModelProviderName.ollama,
+                model_id="qwen3:4b",
+                supports_data_gen=False,
+                reasoning_capable=True,
+                structured_output_mode=StructuredOutputMode.json_schema,
+            ),
+        ],
+    ),
+    # Qwen 3 4B Non-Thinking
+    KilnModel(
+        family=ModelFamily.qwen,
+        name=ModelName.qwen_3_4b_no_thinking,
+        friendly_name="Qwen 3 4B Non-Thinking",
+        providers=[
+            KilnModelProvider(
+                name=ModelProviderName.openrouter,
+                model_id="qwen/qwen3-4b:free",
+                structured_output_mode=StructuredOutputMode.json_instructions,
+                formatter=ModelFormatterID.qwen3_style_no_think,
+                supports_data_gen=False,
+                parser=ModelParserID.optional_r1_thinking,
+            ),
+            KilnModelProvider(
+                name=ModelProviderName.ollama,
+                model_id="qwen3:4b",
+                structured_output_mode=StructuredOutputMode.json_schema,
+                formatter=ModelFormatterID.qwen3_style_no_think,
+                supports_data_gen=False,
+            ),
+        ],
+    ),
+    # Qwen 3 8B
+    KilnModel(
+        family=ModelFamily.qwen,
+        name=ModelName.qwen_3_8b,
+        friendly_name="Qwen 3 8B",
+        providers=[
+            KilnModelProvider(
+                name=ModelProviderName.openrouter,
+                model_id="qwen/qwen3-8b",
+                supports_structured_output=True,
+                structured_output_mode=StructuredOutputMode.json_instructions,
+                reasoning_capable=True,
+                require_openrouter_reasoning=True,
+                r1_openrouter_options=True,
+                parser=ModelParserID.r1_thinking,
+                supports_data_gen=False,
+            ),
+            KilnModelProvider(
+                name=ModelProviderName.ollama,
+                model_id="qwen3:8b",
+                supports_data_gen=False,
+                reasoning_capable=True,
+                structured_output_mode=StructuredOutputMode.json_schema,
+            ),
+        ],
+    ),
+    # Qwen 3 8B Non-Thinking
+    KilnModel(
+        family=ModelFamily.qwen,
+        name=ModelName.qwen_3_8b_no_thinking,
+        friendly_name="Qwen 3 8B Non-Thinking",
+        providers=[
+            KilnModelProvider(
+                name=ModelProviderName.openrouter,
+                model_id="qwen/qwen3-8b",
+                structured_output_mode=StructuredOutputMode.json_instructions,
+                formatter=ModelFormatterID.qwen3_style_no_think,
+                supports_data_gen=False,
+                parser=ModelParserID.optional_r1_thinking,
+            ),
+            KilnModelProvider(
+                name=ModelProviderName.ollama,
+                model_id="qwen3:8b",
+                structured_output_mode=StructuredOutputMode.json_schema,
+                formatter=ModelFormatterID.qwen3_style_no_think,
+                supports_data_gen=False,
+            ),
+        ],
+    ),
+    # Qwen 3 14B
+    KilnModel(
+        family=ModelFamily.qwen,
+        name=ModelName.qwen_3_14b,
+        friendly_name="Qwen 3 14B",
+        providers=[
+            KilnModelProvider(
+                name=ModelProviderName.openrouter,
+                model_id="qwen/qwen3-14b",
+                structured_output_mode=StructuredOutputMode.json_instructions,
+                reasoning_capable=True,
+                require_openrouter_reasoning=True,
+                r1_openrouter_options=True,
+                parser=ModelParserID.r1_thinking,
+                supports_data_gen=True,
+            ),
+            KilnModelProvider(
+                name=ModelProviderName.ollama,
+                model_id="qwen3:14b",
+                supports_data_gen=True,
+                reasoning_capable=True,
+                structured_output_mode=StructuredOutputMode.json_schema,
+            ),
+        ],
+    ),
+    # Qwen 3 14B Non-Thinking
+    KilnModel(
+        family=ModelFamily.qwen,
+        name=ModelName.qwen_3_14b_no_thinking,
+        friendly_name="Qwen 3 14B Non-Thinking",
+        providers=[
+            KilnModelProvider(
+                name=ModelProviderName.openrouter,
+                model_id="qwen/qwen3-14b",
+                structured_output_mode=StructuredOutputMode.json_instructions,
+                formatter=ModelFormatterID.qwen3_style_no_think,
+                supports_data_gen=True,
+                parser=ModelParserID.optional_r1_thinking,
+            ),
+            KilnModelProvider(
+                name=ModelProviderName.ollama,
+                model_id="qwen3:14b",
+                formatter=ModelFormatterID.qwen3_style_no_think,
+                supports_data_gen=True,
+                structured_output_mode=StructuredOutputMode.json_schema,
+            ),
+        ],
+    ),
+    # Qwen 3 30B (3B Active)
+    KilnModel(
+        family=ModelFamily.qwen,
+        name=ModelName.qwen_3_30b_a3b,
+        friendly_name="Qwen 3 30B (3B Active)",
+        providers=[
+            KilnModelProvider(
+                name=ModelProviderName.openrouter,
+                model_id="qwen/qwen3-30b-a3b",
+                structured_output_mode=StructuredOutputMode.json_instructions,
+                reasoning_capable=True,
+                require_openrouter_reasoning=True,
+                r1_openrouter_options=True,
+                parser=ModelParserID.r1_thinking,
+                supports_data_gen=True,
+            ),
+            KilnModelProvider(
+                name=ModelProviderName.ollama,
+                model_id="qwen3:30b-a3b",
+                supports_data_gen=True,
+                reasoning_capable=True,
+                structured_output_mode=StructuredOutputMode.json_schema,
+            ),
+            KilnModelProvider(
+                name=ModelProviderName.fireworks_ai,
+                model_id="accounts/fireworks/models/qwen3-30b-a3b",
+                supports_data_gen=True,
+                reasoning_capable=True,
+                structured_output_mode=StructuredOutputMode.json_instructions,
+                parser=ModelParserID.r1_thinking,
+            ),
+        ],
+    ),
+    # Qwen 3 30B (3B Active) Non-Thinking
+    KilnModel(
+        family=ModelFamily.qwen,
+        name=ModelName.qwen_3_30b_a3b_no_thinking,
+        friendly_name="Qwen 3 30B (3B Active) Non-Thinking",
+        providers=[
+            KilnModelProvider(
+                name=ModelProviderName.openrouter,
+                model_id="qwen/qwen3-30b-a3b",
+                structured_output_mode=StructuredOutputMode.json_instructions,
+                formatter=ModelFormatterID.qwen3_style_no_think,
+                supports_data_gen=True,
+                parser=ModelParserID.optional_r1_thinking,
+            ),
+            KilnModelProvider(
+                name=ModelProviderName.ollama,
+                model_id="qwen3:30b-a3b",
+                structured_output_mode=StructuredOutputMode.json_schema,
+                formatter=ModelFormatterID.qwen3_style_no_think,
+                supports_data_gen=True,
+            ),
+            KilnModelProvider(
+                name=ModelProviderName.fireworks_ai,
+                model_id="accounts/fireworks/models/qwen3-30b-a3b",
+                supports_data_gen=True,
+                formatter=ModelFormatterID.qwen3_style_no_think,
+                structured_output_mode=StructuredOutputMode.json_instructions,
+                parser=ModelParserID.optional_r1_thinking,
+            ),
+        ],
+    ),
+    # Qwen 3 32B
+    KilnModel(
+        family=ModelFamily.qwen,
+        name=ModelName.qwen_3_32b,
+        friendly_name="Qwen 3 32B",
+        providers=[
+            KilnModelProvider(
+                name=ModelProviderName.openrouter,
+                model_id="qwen/qwen3-32b",
+                reasoning_capable=True,
+                require_openrouter_reasoning=True,
+                r1_openrouter_options=True,
+                structured_output_mode=StructuredOutputMode.json_instructions,
+                parser=ModelParserID.r1_thinking,
+                supports_data_gen=True,
+            ),
+            KilnModelProvider(
+                name=ModelProviderName.ollama,
+                model_id="qwen3:32b",
+                supports_data_gen=True,
+                reasoning_capable=True,
+                structured_output_mode=StructuredOutputMode.json_schema,
+            ),
+        ],
+    ),
+    # Qwen 3 32B No Thinking
+    KilnModel(
+        family=ModelFamily.qwen,
+        name=ModelName.qwen_3_32b_no_thinking,
+        friendly_name="Qwen 3 32B Non-Thinking",
+        providers=[
+            KilnModelProvider(
+                name=ModelProviderName.openrouter,
+                model_id="qwen/qwen3-32b",
+                structured_output_mode=StructuredOutputMode.json_instructions,
+                formatter=ModelFormatterID.qwen3_style_no_think,
+                supports_data_gen=True,
+                parser=ModelParserID.optional_r1_thinking,
+            ),
+            KilnModelProvider(
+                name=ModelProviderName.ollama,
+                model_id="qwen3:32b",
+                structured_output_mode=StructuredOutputMode.json_schema,
+                formatter=ModelFormatterID.qwen3_style_no_think,
+                supports_data_gen=True,
+            ),
+        ],
+    ),
+    # Qwen 3 235B (22B Active)
+    KilnModel(
+        family=ModelFamily.qwen,
+        name=ModelName.qwen_3_235b_a22b,
+        friendly_name="Qwen 3 235B (22B Active)",
+        providers=[
+            KilnModelProvider(
+                name=ModelProviderName.openrouter,
+                model_id="qwen/qwen3-235b-a22b",
+                reasoning_capable=True,
+                require_openrouter_reasoning=True,
+                supports_data_gen=True,
+                suggested_for_data_gen=True,
+                r1_openrouter_options=True,
+                structured_output_mode=StructuredOutputMode.json_instructions,
+                parser=ModelParserID.r1_thinking,
+            ),
+            KilnModelProvider(
+                name=ModelProviderName.ollama,
+                model_id="qwen3:235b-a22b",
+                supports_data_gen=True,
+                reasoning_capable=True,
+                structured_output_mode=StructuredOutputMode.json_schema,
+            ),
+            KilnModelProvider(
+                name=ModelProviderName.fireworks_ai,
+                model_id="accounts/fireworks/models/qwen3-235b-a22b",
+                supports_data_gen=True,
+                reasoning_capable=True,
+                structured_output_mode=StructuredOutputMode.json_instructions,
+                parser=ModelParserID.r1_thinking,
+            ),
+            KilnModelProvider(
+                name=ModelProviderName.together_ai,
+                model_id="Qwen/Qwen3-235B-A22B-fp8-tput",
+                supports_data_gen=True,
+                reasoning_capable=True,
+                structured_output_mode=StructuredOutputMode.json_instructions,
+                parser=ModelParserID.r1_thinking,
+            ),
+        ],
+    ),
+    # Qwen 3 235B (22B Active) Non-Thinking
+    KilnModel(
+        family=ModelFamily.qwen,
+        name=ModelName.qwen_3_235b_a22b_no_thinking,
+        friendly_name="Qwen 3 235B (22B Active) Non-Thinking",
+        providers=[
+            KilnModelProvider(
+                name=ModelProviderName.openrouter,
+                model_id="qwen/qwen3-235b-a22b",
+                structured_output_mode=StructuredOutputMode.json_instructions,
+                formatter=ModelFormatterID.qwen3_style_no_think,
+                supports_data_gen=True,
+                reasoning_capable=False,
+                parser=ModelParserID.optional_r1_thinking,
+            ),
+            KilnModelProvider(
+                name=ModelProviderName.ollama,
+                model_id="qwen3:235b-a22b",
+                structured_output_mode=StructuredOutputMode.json_schema,
+                formatter=ModelFormatterID.qwen3_style_no_think,
+                supports_data_gen=True,
+            ),
+            KilnModelProvider(
+                name=ModelProviderName.fireworks_ai,
+                model_id="accounts/fireworks/models/qwen3-235b-a22b",
+                supports_data_gen=True,
+                formatter=ModelFormatterID.qwen3_style_no_think,
+                structured_output_mode=StructuredOutputMode.json_instructions,
+                parser=ModelParserID.optional_r1_thinking,
+            ),
+            KilnModelProvider(
+                name=ModelProviderName.together_ai,
+                model_id="Qwen/Qwen3-235B-A22B-fp8-tput",
+                supports_data_gen=True,
+                formatter=ModelFormatterID.qwen3_style_no_think,
+                structured_output_mode=StructuredOutputMode.json_instructions,
+                parser=ModelParserID.optional_r1_thinking,
+            ),
+        ],
+    ),
 ]
+
+
+def get_model_by_name(name: ModelName) -> KilnModel:
+    for model in built_in_models:
+        if model.name == name:
+            return model
+    raise ValueError(f"Model {name} not found in the list of built-in models")
+
+
+def default_structured_output_mode_for_model_provider(
+    model_name: str,
+    provider: ModelProviderName,
+    default: StructuredOutputMode = StructuredOutputMode.default,
+    disallowed_modes: List[StructuredOutputMode] = [],
+) -> StructuredOutputMode:
+    """
+    We don't expose setting this manually in the UI, so pull a recommended mode from ml_model_list
+    """
+    try:
+        # Convert string to ModelName enum
+        model_name_enum = ModelName(model_name)
+        model = get_model_by_name(model_name_enum)
+    except (ValueError, KeyError):
+        # If model not found, return default
+        return default
+
+    # Find the provider within the model's providers
+    for model_provider in model.providers:
+        if model_provider.name == provider:
+            mode = model_provider.structured_output_mode
+            if mode not in disallowed_modes:
+                return mode
+
+    # If provider not found, return default
+    return default