diff --git a/models/alibaba/pre-mock_new_gateway_session_affinity.yaml b/models/alibaba/pre-mock_new_gateway_session_affinity.yaml new file mode 100644 index 00000000..79911082 --- /dev/null +++ b/models/alibaba/pre-mock_new_gateway_session_affinity.yaml @@ -0,0 +1,55 @@ +# yaml-language-server: $schema=https://modelparams.dev/api/v1/schema.json +provider: alibaba +authType: api_key +apiSurface: openai-chat-completions +model: pre-mock_new_gateway_session_affinity +params: + - path: max_completion_tokens + type: integer + label: Max tokens + description: Maximum number of tokens to generate, including both reasoning and the final answer. + range: + min: 1 + group: generation_length + - path: temperature + type: number + label: Temperature + description: Controls randomness. Lower values make outputs more focused; higher values make them more varied. + range: + min: 0 + max: 1.9 + step: 0.1 + group: sampling + - path: top_p + type: number + label: Top P + description: Controls nucleus sampling by limiting generation to tokens within the selected cumulative probability. + range: + min: 0 + max: 1 + step: 0.01 + group: sampling + - path: extra_body.top_k + type: integer + label: Top K + description: Limits generation to the selected number of highest-probability tokens. Values above 100 disable top-k sampling. + default: 20 + range: + min: 0 + group: sampling + - path: extra_body.enable_thinking + type: boolean + label: Enable thinking + description: Toggles the model's hybrid thinking mode, sent as a provider-specific extra body field on OpenAI-compatible clients. + default: true + group: reasoning + - path: extra_body.thinking_budget + type: integer + label: Thinking budget + description: Maximum number of tokens the model may spend on reasoning before it starts the final answer; defaults to the model's maximum reasoning length. + range: + min: 1 + group: reasoning + applicability: + only: + extra_body.enable_thinking: true diff --git a/models/fireworks/muse-spark-open.yaml b/models/fireworks/muse-spark-open.yaml new file mode 100644 index 00000000..b31849ee --- /dev/null +++ b/models/fireworks/muse-spark-open.yaml @@ -0,0 +1,66 @@ +# yaml-language-server: $schema=https://modelparams.dev/api/v1/schema.json +provider: fireworks +authType: api_key +apiSurface: openai-chat-completions +model: muse-spark-open +wireId: accounts/fireworks/models/muse-spark-open +params: + - path: max_tokens + type: integer + label: Max tokens + description: Maximum number of tokens to generate in the response. + range: + min: 1 + group: generation_length + - path: temperature + type: number + label: Temperature + description: Controls randomness. Lower values make outputs more focused; higher values make them more varied. + range: + min: 0 + max: 2 + step: 0.1 + group: sampling + - path: top_p + type: number + label: Top P + description: Controls nucleus sampling by limiting generation to tokens within the selected cumulative probability. + range: + min: 0 + max: 1 + step: 0.01 + group: sampling + - path: top_k + type: integer + label: Top K + description: Limits token sampling to the top K most likely next tokens. + range: + min: 1 + max: 100 + group: sampling + - path: presence_penalty + type: number + label: Presence penalty + description: Penalizes tokens that have already appeared to encourage a wider variety of content. + range: + min: -2 + max: 2 + step: 0.1 + group: sampling + - path: frequency_penalty + type: number + label: Frequency penalty + description: Penalizes tokens in proportion to how often they have already appeared. + range: + min: -2 + max: 2 + step: 0.1 + group: sampling + - path: response_format.type + type: enum + label: Response format + description: Choose plain text or force valid JSON output. + values: + - text + - json_object + group: output_format diff --git a/models/nvidia/nemotron-3.5-super-vl-120b-a12b.yaml b/models/nvidia/nemotron-3.5-super-vl-120b-a12b.yaml new file mode 100644 index 00000000..92f0ac9e --- /dev/null +++ b/models/nvidia/nemotron-3.5-super-vl-120b-a12b.yaml @@ -0,0 +1,51 @@ +# yaml-language-server: $schema=https://modelparams.dev/api/v1/schema.json +provider: nvidia +authType: api_key +apiSurface: openai-chat-completions +model: nemotron-3.5-super-vl-120b-a12b +wireId: nvidia/nemotron-3.5-super-vl-120b-a12b +params: + - path: temperature + type: number + label: Temperature + description: Controls randomness. Lower values make outputs more focused; higher values make them more varied. Not recommended to modify both temperature and top_p in the same call. + default: 1 + range: + max: 1 + group: sampling + - path: top_p + type: number + label: Top P + description: Controls nucleus sampling by limiting generation to tokens within the selected cumulative probability. Not recommended to modify both temperature and top_p in the same call. + default: 0.95 + range: + max: 1 + group: sampling + - path: max_tokens + type: integer + label: Max tokens + description: Maximum number of tokens to generate. Generation stops when this limit is reached. + default: 16384 + range: + min: 1 + max: 32768 + group: generation_length + - path: reasoning_effort + type: enum + label: Reasoning effort + description: Controls the reasoning mode. 'none' disables reasoning tokens, 'low' enables low-effort reasoning, and 'high' enables full reasoning. + default: high + values: + - none + - low + - high + group: reasoning + - path: reasoning_budget + type: integer + label: Reasoning budget + description: Maximum number of tokens the model may use for internal reasoning before being forced to end the reasoning trace. Use -1 to disable budget enforcement. + default: 16384 + range: + min: -1 + max: 32768 + group: reasoning diff --git a/models/nvidia/step-5-preview.yaml b/models/nvidia/step-5-preview.yaml new file mode 100644 index 00000000..f3755351 --- /dev/null +++ b/models/nvidia/step-5-preview.yaml @@ -0,0 +1,51 @@ +# yaml-language-server: $schema=https://modelparams.dev/api/v1/schema.json +provider: nvidia +authType: api_key +apiSurface: openai-chat-completions +model: step-5-preview +wireId: stepfun-ai/step-5-preview +params: + - path: temperature + type: number + label: Temperature + description: Controls randomness. Lower values make outputs more focused; higher values make them more varied. Not recommended to modify both temperature and top_p in the same call. + default: 0.6 + range: + min: 0 + max: 1 + group: sampling + - path: top_p + type: number + label: Top P + description: Controls nucleus sampling by limiting generation to tokens within the selected cumulative probability. Not recommended to modify both temperature and top_p in the same call. + default: 0.95 + range: + max: 1 + group: sampling + - path: max_tokens + type: integer + label: Max tokens + description: Maximum number of tokens to generate. Generation stops when this limit is reached. + default: 4096 + range: + min: 1 + max: 16384 + group: generation_length + - path: frequency_penalty + type: number + label: Frequency penalty + description: Penalizes new tokens based on their existing frequency in the text so far, decreasing the model's likelihood to repeat the same line verbatim. + default: 0 + range: + min: -2 + max: 2 + group: sampling + - path: presence_penalty + type: number + label: Presence penalty + description: 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. + default: 0 + range: + min: -2 + max: 2 + group: sampling diff --git a/models/xai/grok-4.7.yaml b/models/xai/grok-4.7.yaml index 466703d1..c50d8fd1 100644 --- a/models/xai/grok-4.7.yaml +++ b/models/xai/grok-4.7.yaml @@ -28,7 +28,7 @@ params: description: Controls nucleus sampling by limiting generation to tokens within the selected cumulative probability. default: 1 range: - min: 0 + min: 0.01 max: 1 step: 0.01 group: sampling diff --git a/packages/modelparams-python/src/modelparams/_generated/catalog.json b/packages/modelparams-python/src/modelparams/_generated/catalog.json index f84cf7c8..5168dae4 100644 --- a/packages/modelparams-python/src/modelparams/_generated/catalog.json +++ b/packages/modelparams-python/src/modelparams/_generated/catalog.json @@ -987,6 +987,82 @@ } ] }, + { + "provider": "alibaba", + "authType": "api_key", + "apiSurface": "openai-chat-completions", + "model": "pre-mock_new_gateway_session_affinity", + "params": [ + { + "path": "max_completion_tokens", + "label": "Max tokens", + "description": "Maximum number of tokens to generate, including both reasoning and the final answer.", + "group": "generation_length", + "type": "integer", + "range": { + "min": 1 + } + }, + { + "path": "temperature", + "label": "Temperature", + "description": "Controls randomness. Lower values make outputs more focused; higher values make them more varied.", + "group": "sampling", + "type": "number", + "range": { + "min": 0, + "max": 1.9, + "step": 0.1 + } + }, + { + "path": "top_p", + "label": "Top P", + "description": "Controls nucleus sampling by limiting generation to tokens within the selected cumulative probability.", + "group": "sampling", + "type": "number", + "range": { + "min": 0, + "max": 1, + "step": 0.01 + } + }, + { + "path": "extra_body.top_k", + "label": "Top K", + "description": "Limits generation to the selected number of highest-probability tokens. Values above 100 disable top-k sampling.", + "group": "sampling", + "type": "integer", + "default": 20, + "range": { + "min": 0 + } + }, + { + "path": "extra_body.enable_thinking", + "label": "Enable thinking", + "description": "Toggles the model's hybrid thinking mode, sent as a provider-specific extra body field on OpenAI-compatible clients.", + "group": "reasoning", + "type": "boolean", + "default": true + }, + { + "path": "extra_body.thinking_budget", + "label": "Thinking budget", + "description": "Maximum number of tokens the model may spend on reasoning before it starts the final answer; defaults to the model's maximum reasoning length.", + "group": "reasoning", + "applicability": { + "only": { + "extra_body.enable_thinking": true + } + }, + "type": "integer", + "range": { + "min": 1 + } + } + ] + }, { "provider": "alibaba", "authType": "api_key", @@ -16336,6 +16412,95 @@ } ] }, + { + "provider": "fireworks", + "authType": "api_key", + "apiSurface": "openai-chat-completions", + "model": "muse-spark-open", + "wireId": "accounts/fireworks/models/muse-spark-open", + "params": [ + { + "path": "max_tokens", + "label": "Max tokens", + "description": "Maximum number of tokens to generate in the response.", + "group": "generation_length", + "type": "integer", + "range": { + "min": 1 + } + }, + { + "path": "temperature", + "label": "Temperature", + "description": "Controls randomness. Lower values make outputs more focused; higher values make them more varied.", + "group": "sampling", + "type": "number", + "range": { + "min": 0, + "max": 2, + "step": 0.1 + } + }, + { + "path": "top_p", + "label": "Top P", + "description": "Controls nucleus sampling by limiting generation to tokens within the selected cumulative probability.", + "group": "sampling", + "type": "number", + "range": { + "min": 0, + "max": 1, + "step": 0.01 + } + }, + { + "path": "top_k", + "label": "Top K", + "description": "Limits token sampling to the top K most likely next tokens.", + "group": "sampling", + "type": "integer", + "range": { + "min": 1, + "max": 100 + } + }, + { + "path": "presence_penalty", + "label": "Presence penalty", + "description": "Penalizes tokens that have already appeared to encourage a wider variety of content.", + "group": "sampling", + "type": "number", + "range": { + "min": -2, + "max": 2, + "step": 0.1 + } + }, + { + "path": "frequency_penalty", + "label": "Frequency penalty", + "description": "Penalizes tokens in proportion to how often they have already appeared.", + "group": "sampling", + "type": "number", + "range": { + "min": -2, + "max": 2, + "step": 0.1 + } + }, + { + "path": "response_format.type", + "label": "Response format", + "description": "Choose plain text or force valid JSON output.", + "group": "output_format", + "type": "enum", + "values": [ + "text", + "json_object" + ] + } + ] + }, { "provider": "fireworks", "authType": "api_key", @@ -28015,6 +28180,74 @@ } ] }, + { + "provider": "nvidia", + "authType": "api_key", + "apiSurface": "openai-chat-completions", + "model": "nemotron-3.5-super-vl-120b-a12b", + "wireId": "nvidia/nemotron-3.5-super-vl-120b-a12b", + "params": [ + { + "path": "temperature", + "label": "Temperature", + "description": "Controls randomness. Lower values make outputs more focused; higher values make them more varied. Not recommended to modify both temperature and top_p in the same call.", + "group": "sampling", + "type": "number", + "default": 1, + "range": { + "max": 1 + } + }, + { + "path": "top_p", + "label": "Top P", + "description": "Controls nucleus sampling by limiting generation to tokens within the selected cumulative probability. Not recommended to modify both temperature and top_p in the same call.", + "group": "sampling", + "type": "number", + "default": 0.95, + "range": { + "max": 1 + } + }, + { + "path": "max_tokens", + "label": "Max tokens", + "description": "Maximum number of tokens to generate. Generation stops when this limit is reached.", + "group": "generation_length", + "type": "integer", + "default": 16384, + "range": { + "min": 1, + "max": 32768 + } + }, + { + "path": "reasoning_effort", + "label": "Reasoning effort", + "description": "Controls the reasoning mode. 'none' disables reasoning tokens, 'low' enables low-effort reasoning, and 'high' enables full reasoning.", + "group": "reasoning", + "type": "enum", + "default": "high", + "values": [ + "none", + "low", + "high" + ] + }, + { + "path": "reasoning_budget", + "label": "Reasoning budget", + "description": "Maximum number of tokens the model may use for internal reasoning before being forced to end the reasoning trace. Use -1 to disable budget enforcement.", + "group": "reasoning", + "type": "integer", + "default": 16384, + "range": { + "min": -1, + "max": 32768 + } + } + ] + }, { "provider": "nvidia", "authType": "api_key", @@ -28223,6 +28456,74 @@ } ] }, + { + "provider": "nvidia", + "authType": "api_key", + "apiSurface": "openai-chat-completions", + "model": "step-5-preview", + "wireId": "stepfun-ai/step-5-preview", + "params": [ + { + "path": "temperature", + "label": "Temperature", + "description": "Controls randomness. Lower values make outputs more focused; higher values make them more varied. Not recommended to modify both temperature and top_p in the same call.", + "group": "sampling", + "type": "number", + "default": 0.6, + "range": { + "min": 0, + "max": 1 + } + }, + { + "path": "top_p", + "label": "Top P", + "description": "Controls nucleus sampling by limiting generation to tokens within the selected cumulative probability. Not recommended to modify both temperature and top_p in the same call.", + "group": "sampling", + "type": "number", + "default": 0.95, + "range": { + "max": 1 + } + }, + { + "path": "max_tokens", + "label": "Max tokens", + "description": "Maximum number of tokens to generate. Generation stops when this limit is reached.", + "group": "generation_length", + "type": "integer", + "default": 4096, + "range": { + "min": 1, + "max": 16384 + } + }, + { + "path": "frequency_penalty", + "label": "Frequency penalty", + "description": "Penalizes new tokens based on their existing frequency in the text so far, decreasing the model's likelihood to repeat the same line verbatim.", + "group": "sampling", + "type": "number", + "default": 0, + "range": { + "min": -2, + "max": 2 + } + }, + { + "path": "presence_penalty", + "label": "Presence penalty", + "description": "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.", + "group": "sampling", + "type": "number", + "default": 0, + "range": { + "min": -2, + "max": 2 + } + } + ] + }, { "provider": "nvidia", "authType": "api_key", @@ -35579,7 +35880,7 @@ "type": "number", "default": 1, "range": { - "min": 0, + "min": 0.01, "max": 1, "step": 0.01 } diff --git a/packages/modelparams-python/src/modelparams/_generated/model_ids.py b/packages/modelparams-python/src/modelparams/_generated/model_ids.py index 970579ff..ef44d636 100644 --- a/packages/modelparams-python/src/modelparams/_generated/model_ids.py +++ b/packages/modelparams-python/src/modelparams/_generated/model_ids.py @@ -17,6 +17,7 @@ "alibaba/glm-5.3-prime", "alibaba/kimi-k2.7-code", "alibaba/kimi-k3", + "alibaba/pre-mock_new_gateway_session_affinity", "alibaba/qvq-max", "alibaba/qwen-coder-plus", "alibaba/qwen-flash", @@ -216,6 +217,7 @@ "fireworks/minimax-m2p7", "fireworks/minimax-m3", "fireworks/muse-glimmer-30b", + "fireworks/muse-spark-open", "fireworks/nemotron-3-ultra-nvfp4", "fireworks/nemotron-lightning-3p5-30b-a3b", "fireworks/qwen3p7-plus", @@ -354,9 +356,11 @@ "nvidia/nemotron-3-ultra-550b-a55b", "nvidia/nemotron-3-ultra-subscription", "nvidia/nemotron-3.5-lightning-30b-a3b", + "nvidia/nemotron-3.5-super-vl-120b-a12b", "nvidia/nemotron-content-safety-reasoning-4b", "nvidia/nemotron-mini-4b-instruct", "nvidia/riva-translate-4b-instruct-v1.1", + "nvidia/step-5-preview", "nvidia/usdcode-llama-3.1-70b-instruct", "openai/chatgpt-4o-latest", "openai/gpt-3.5-turbo", @@ -511,6 +515,7 @@ "alibaba/glm-5.3-prime", "alibaba/kimi-k2.7-code", "alibaba/kimi-k3", + "alibaba/pre-mock_new_gateway_session_affinity", "alibaba/qvq-max", "alibaba/qwen-coder-plus", "alibaba/qwen-flash", @@ -710,6 +715,7 @@ "fireworks/minimax-m2p7", "fireworks/minimax-m3", "fireworks/muse-glimmer-30b", + "fireworks/muse-spark-open", "fireworks/nemotron-3-ultra-nvfp4", "fireworks/nemotron-lightning-3p5-30b-a3b", "fireworks/qwen3p7-plus", @@ -848,9 +854,11 @@ "nvidia/nemotron-3-ultra-550b-a55b", "nvidia/nemotron-3-ultra-subscription", "nvidia/nemotron-3.5-lightning-30b-a3b", + "nvidia/nemotron-3.5-super-vl-120b-a12b", "nvidia/nemotron-content-safety-reasoning-4b", "nvidia/nemotron-mini-4b-instruct", "nvidia/riva-translate-4b-instruct-v1.1", + "nvidia/step-5-preview", "nvidia/usdcode-llama-3.1-70b-instruct", "openai/chatgpt-4o-latest", "openai/gpt-3.5-turbo", diff --git a/packages/modelparams-python/src/modelparams/_generated/registry.py b/packages/modelparams-python/src/modelparams/_generated/registry.py index ede7842c..e7d7465e 100644 --- a/packages/modelparams-python/src/modelparams/_generated/registry.py +++ b/packages/modelparams-python/src/modelparams/_generated/registry.py @@ -42,6 +42,7 @@ "alibaba/glm-5.3-prime": alibaba.Glm_5_3_PrimeParams, "alibaba/kimi-k2.7-code": alibaba.Kimi_K2_7_CodeParams, "alibaba/kimi-k3": alibaba.Kimi_K3Params, + "alibaba/pre-mock_new_gateway_session_affinity": alibaba.Pre_Mock_New_Gateway_Session_AffinityParams, "alibaba/qvq-max": alibaba.Qvq_MaxParams, "alibaba/qwen-coder-plus": alibaba.Qwen_Coder_PlusParams, "alibaba/qwen-flash": alibaba.Qwen_FlashParams, @@ -241,6 +242,7 @@ "fireworks/minimax-m2p7": fireworks.Minimax_M2p7Params, "fireworks/minimax-m3": fireworks.Minimax_M3Params, "fireworks/muse-glimmer-30b": fireworks.Muse_Glimmer_30bParams, + "fireworks/muse-spark-open": fireworks.Muse_Spark_OpenParams, "fireworks/nemotron-3-ultra-nvfp4": fireworks.Nemotron_3_Ultra_Nvfp4Params, "fireworks/nemotron-lightning-3p5-30b-a3b": fireworks.Nemotron_Lightning_3p5_30b_A3bParams, "fireworks/qwen3p7-plus": fireworks.Qwen3p7_PlusParams, @@ -379,9 +381,11 @@ "nvidia/nemotron-3-ultra-550b-a55b": nvidia.Nemotron_3_Ultra_550b_A55bParams, "nvidia/nemotron-3-ultra-subscription": nvidia.Nemotron_3_Ultra_SubscriptionParams, "nvidia/nemotron-3.5-lightning-30b-a3b": nvidia.Nemotron_3_5_Lightning_30b_A3bParams, + "nvidia/nemotron-3.5-super-vl-120b-a12b": nvidia.Nemotron_3_5_Super_Vl_120b_A12bParams, "nvidia/nemotron-content-safety-reasoning-4b": nvidia.Nemotron_Content_Safety_Reasoning_4bParams, "nvidia/nemotron-mini-4b-instruct": nvidia.Nemotron_Mini_4b_InstructParams, "nvidia/riva-translate-4b-instruct-v1.1": nvidia.Riva_Translate_4b_Instruct_V1_1Params, + "nvidia/step-5-preview": nvidia.Step_5_PreviewParams, "nvidia/usdcode-llama-3.1-70b-instruct": nvidia.Usdcode_Llama_3_1_70b_InstructParams, "openai/chatgpt-4o-latest": openai.Chatgpt_4o_LatestParams, "openai/gpt-3.5-turbo": openai.Gpt_3_5_TurboParams, diff --git a/packages/modelparams-python/src/modelparams/types/alibaba.py b/packages/modelparams-python/src/modelparams/types/alibaba.py index 444f1e83..cf5323dc 100644 --- a/packages/modelparams-python/src/modelparams/types/alibaba.py +++ b/packages/modelparams-python/src/modelparams/types/alibaba.py @@ -192,6 +192,20 @@ ) setattr(Kimi_K3Params, "__pydantic_config__", _PARAMS_CONFIG) +Pre_Mock_New_Gateway_Session_AffinityParams = TypedDict( + "Pre_Mock_New_Gateway_Session_AffinityParams", + { + "max_completion_tokens": Annotated[int, Field(ge=1)], + "temperature": Annotated[float, Field(ge=0, le=1.9)], + "top_p": Annotated[float, Field(ge=0, le=1)], + "extra_body.top_k": Annotated[int, Field(ge=0)], + "extra_body.enable_thinking": bool, + "extra_body.thinking_budget": Annotated[int, Field(ge=1)], + }, + total=False, +) +setattr(Pre_Mock_New_Gateway_Session_AffinityParams, "__pydantic_config__", _PARAMS_CONFIG) + Qvq_MaxParams = TypedDict( "Qvq_MaxParams", { @@ -1075,6 +1089,7 @@ "Glm_5_3_PrimeParams", "Kimi_K2_7_CodeParams", "Kimi_K3Params", + "Pre_Mock_New_Gateway_Session_AffinityParams", "Qvq_MaxParams", "Qwen_Coder_PlusParams", "Qwen_FlashParams", diff --git a/packages/modelparams-python/src/modelparams/types/fireworks.py b/packages/modelparams-python/src/modelparams/types/fireworks.py index 5be22d0c..e78e66d5 100644 --- a/packages/modelparams-python/src/modelparams/types/fireworks.py +++ b/packages/modelparams-python/src/modelparams/types/fireworks.py @@ -280,6 +280,21 @@ ) setattr(Muse_Glimmer_30bParams, "__pydantic_config__", _PARAMS_CONFIG) +Muse_Spark_OpenParams = TypedDict( + "Muse_Spark_OpenParams", + { + "max_tokens": Annotated[int, Field(ge=1)], + "temperature": Annotated[float, Field(ge=0, le=2)], + "top_p": Annotated[float, Field(ge=0, le=1)], + "top_k": Annotated[int, Field(ge=1, le=100)], + "presence_penalty": Annotated[float, Field(ge=-2, le=2)], + "frequency_penalty": Annotated[float, Field(ge=-2, le=2)], + "response_format.type": Literal["text", "json_object"], + }, + total=False, +) +setattr(Muse_Spark_OpenParams, "__pydantic_config__", _PARAMS_CONFIG) + Nemotron_3_Ultra_Nvfp4Params = TypedDict( "Nemotron_3_Ultra_Nvfp4Params", { @@ -374,6 +389,7 @@ "Minimax_M2p7Params", "Minimax_M3Params", "Muse_Glimmer_30bParams", + "Muse_Spark_OpenParams", "Nemotron_3_Ultra_Nvfp4Params", "Nemotron_Lightning_3p5_30b_A3bParams", "Qwen3p7_PlusParams", diff --git a/packages/modelparams-python/src/modelparams/types/nvidia.py b/packages/modelparams-python/src/modelparams/types/nvidia.py index 08f2fb7b..c0ec3c74 100644 --- a/packages/modelparams-python/src/modelparams/types/nvidia.py +++ b/packages/modelparams-python/src/modelparams/types/nvidia.py @@ -397,6 +397,19 @@ ) setattr(Nemotron_3_5_Lightning_30b_A3bParams, "__pydantic_config__", _PARAMS_CONFIG) +Nemotron_3_5_Super_Vl_120b_A12bParams = TypedDict( + "Nemotron_3_5_Super_Vl_120b_A12bParams", + { + "temperature": Annotated[float, Field(le=1)], + "top_p": Annotated[float, Field(le=1)], + "max_tokens": Annotated[int, Field(ge=1, le=32768)], + "reasoning_effort": Literal["none", "low", "high"], + "reasoning_budget": Annotated[int, Field(ge=-1, le=32768)], + }, + total=False, +) +setattr(Nemotron_3_5_Super_Vl_120b_A12bParams, "__pydantic_config__", _PARAMS_CONFIG) + Nemotron_Content_Safety_Reasoning_4bParams = TypedDict( "Nemotron_Content_Safety_Reasoning_4bParams", { @@ -438,6 +451,19 @@ ) setattr(Riva_Translate_4b_Instruct_V1_1Params, "__pydantic_config__", _PARAMS_CONFIG) +Step_5_PreviewParams = TypedDict( + "Step_5_PreviewParams", + { + "temperature": Annotated[float, Field(ge=0, le=1)], + "top_p": Annotated[float, Field(le=1)], + "max_tokens": Annotated[int, Field(ge=1, le=16384)], + "frequency_penalty": Annotated[float, Field(ge=-2, le=2)], + "presence_penalty": Annotated[float, Field(ge=-2, le=2)], + }, + total=False, +) +setattr(Step_5_PreviewParams, "__pydantic_config__", _PARAMS_CONFIG) + Usdcode_Llama_3_1_70b_InstructParams = TypedDict( "Usdcode_Llama_3_1_70b_InstructParams", { @@ -479,8 +505,10 @@ "Nemotron_3_Ultra_550b_A55bParams", "Nemotron_3_Ultra_SubscriptionParams", "Nemotron_3_5_Lightning_30b_A3bParams", + "Nemotron_3_5_Super_Vl_120b_A12bParams", "Nemotron_Content_Safety_Reasoning_4bParams", "Nemotron_Mini_4b_InstructParams", "Riva_Translate_4b_Instruct_V1_1Params", + "Step_5_PreviewParams", "Usdcode_Llama_3_1_70b_InstructParams", ] diff --git a/packages/modelparams-python/src/modelparams/types/xai.py b/packages/modelparams-python/src/modelparams/types/xai.py index b6083efa..4d9b70e3 100644 --- a/packages/modelparams-python/src/modelparams/types/xai.py +++ b/packages/modelparams-python/src/modelparams/types/xai.py @@ -167,7 +167,7 @@ { "max_completion_tokens": Annotated[int, Field(ge=1)], "temperature": Annotated[float, Field(ge=0, le=2)], - "top_p": Annotated[float, Field(ge=0, le=1)], + "top_p": Annotated[float, Field(ge=0.01, le=1)], "seed": int, "stop": str, "reasoning_effort": Literal["low", "medium", "high", "xhigh"], diff --git a/packages/modelparams/src/generated/data.ts b/packages/modelparams/src/generated/data.ts index 2cd83d3d..01a1ffa1 100644 --- a/packages/modelparams/src/generated/data.ts +++ b/packages/modelparams/src/generated/data.ts @@ -992,6 +992,82 @@ const GENERATED_CATALOG = [ } ] }, + { + "provider": "alibaba", + "authType": "api_key", + "apiSurface": "openai-chat-completions", + "model": "pre-mock_new_gateway_session_affinity", + "params": [ + { + "path": "max_completion_tokens", + "label": "Max tokens", + "description": "Maximum number of tokens to generate, including both reasoning and the final answer.", + "group": "generation_length", + "type": "integer", + "range": { + "min": 1 + } + }, + { + "path": "temperature", + "label": "Temperature", + "description": "Controls randomness. Lower values make outputs more focused; higher values make them more varied.", + "group": "sampling", + "type": "number", + "range": { + "min": 0, + "max": 1.9, + "step": 0.1 + } + }, + { + "path": "top_p", + "label": "Top P", + "description": "Controls nucleus sampling by limiting generation to tokens within the selected cumulative probability.", + "group": "sampling", + "type": "number", + "range": { + "min": 0, + "max": 1, + "step": 0.01 + } + }, + { + "path": "extra_body.top_k", + "label": "Top K", + "description": "Limits generation to the selected number of highest-probability tokens. Values above 100 disable top-k sampling.", + "group": "sampling", + "type": "integer", + "default": 20, + "range": { + "min": 0 + } + }, + { + "path": "extra_body.enable_thinking", + "label": "Enable thinking", + "description": "Toggles the model's hybrid thinking mode, sent as a provider-specific extra body field on OpenAI-compatible clients.", + "group": "reasoning", + "type": "boolean", + "default": true + }, + { + "path": "extra_body.thinking_budget", + "label": "Thinking budget", + "description": "Maximum number of tokens the model may spend on reasoning before it starts the final answer; defaults to the model's maximum reasoning length.", + "group": "reasoning", + "applicability": { + "only": { + "extra_body.enable_thinking": true + } + }, + "type": "integer", + "range": { + "min": 1 + } + } + ] + }, { "provider": "alibaba", "authType": "api_key", @@ -16341,6 +16417,95 @@ const GENERATED_CATALOG = [ } ] }, + { + "provider": "fireworks", + "authType": "api_key", + "apiSurface": "openai-chat-completions", + "model": "muse-spark-open", + "wireId": "accounts/fireworks/models/muse-spark-open", + "params": [ + { + "path": "max_tokens", + "label": "Max tokens", + "description": "Maximum number of tokens to generate in the response.", + "group": "generation_length", + "type": "integer", + "range": { + "min": 1 + } + }, + { + "path": "temperature", + "label": "Temperature", + "description": "Controls randomness. Lower values make outputs more focused; higher values make them more varied.", + "group": "sampling", + "type": "number", + "range": { + "min": 0, + "max": 2, + "step": 0.1 + } + }, + { + "path": "top_p", + "label": "Top P", + "description": "Controls nucleus sampling by limiting generation to tokens within the selected cumulative probability.", + "group": "sampling", + "type": "number", + "range": { + "min": 0, + "max": 1, + "step": 0.01 + } + }, + { + "path": "top_k", + "label": "Top K", + "description": "Limits token sampling to the top K most likely next tokens.", + "group": "sampling", + "type": "integer", + "range": { + "min": 1, + "max": 100 + } + }, + { + "path": "presence_penalty", + "label": "Presence penalty", + "description": "Penalizes tokens that have already appeared to encourage a wider variety of content.", + "group": "sampling", + "type": "number", + "range": { + "min": -2, + "max": 2, + "step": 0.1 + } + }, + { + "path": "frequency_penalty", + "label": "Frequency penalty", + "description": "Penalizes tokens in proportion to how often they have already appeared.", + "group": "sampling", + "type": "number", + "range": { + "min": -2, + "max": 2, + "step": 0.1 + } + }, + { + "path": "response_format.type", + "label": "Response format", + "description": "Choose plain text or force valid JSON output.", + "group": "output_format", + "type": "enum", + "values": [ + "text", + "json_object" + ] + } + ] + }, { "provider": "fireworks", "authType": "api_key", @@ -28020,6 +28185,74 @@ const GENERATED_CATALOG = [ } ] }, + { + "provider": "nvidia", + "authType": "api_key", + "apiSurface": "openai-chat-completions", + "model": "nemotron-3.5-super-vl-120b-a12b", + "wireId": "nvidia/nemotron-3.5-super-vl-120b-a12b", + "params": [ + { + "path": "temperature", + "label": "Temperature", + "description": "Controls randomness. Lower values make outputs more focused; higher values make them more varied. Not recommended to modify both temperature and top_p in the same call.", + "group": "sampling", + "type": "number", + "default": 1, + "range": { + "max": 1 + } + }, + { + "path": "top_p", + "label": "Top P", + "description": "Controls nucleus sampling by limiting generation to tokens within the selected cumulative probability. Not recommended to modify both temperature and top_p in the same call.", + "group": "sampling", + "type": "number", + "default": 0.95, + "range": { + "max": 1 + } + }, + { + "path": "max_tokens", + "label": "Max tokens", + "description": "Maximum number of tokens to generate. Generation stops when this limit is reached.", + "group": "generation_length", + "type": "integer", + "default": 16384, + "range": { + "min": 1, + "max": 32768 + } + }, + { + "path": "reasoning_effort", + "label": "Reasoning effort", + "description": "Controls the reasoning mode. 'none' disables reasoning tokens, 'low' enables low-effort reasoning, and 'high' enables full reasoning.", + "group": "reasoning", + "type": "enum", + "default": "high", + "values": [ + "none", + "low", + "high" + ] + }, + { + "path": "reasoning_budget", + "label": "Reasoning budget", + "description": "Maximum number of tokens the model may use for internal reasoning before being forced to end the reasoning trace. Use -1 to disable budget enforcement.", + "group": "reasoning", + "type": "integer", + "default": 16384, + "range": { + "min": -1, + "max": 32768 + } + } + ] + }, { "provider": "nvidia", "authType": "api_key", @@ -28228,6 +28461,74 @@ const GENERATED_CATALOG = [ } ] }, + { + "provider": "nvidia", + "authType": "api_key", + "apiSurface": "openai-chat-completions", + "model": "step-5-preview", + "wireId": "stepfun-ai/step-5-preview", + "params": [ + { + "path": "temperature", + "label": "Temperature", + "description": "Controls randomness. Lower values make outputs more focused; higher values make them more varied. Not recommended to modify both temperature and top_p in the same call.", + "group": "sampling", + "type": "number", + "default": 0.6, + "range": { + "min": 0, + "max": 1 + } + }, + { + "path": "top_p", + "label": "Top P", + "description": "Controls nucleus sampling by limiting generation to tokens within the selected cumulative probability. Not recommended to modify both temperature and top_p in the same call.", + "group": "sampling", + "type": "number", + "default": 0.95, + "range": { + "max": 1 + } + }, + { + "path": "max_tokens", + "label": "Max tokens", + "description": "Maximum number of tokens to generate. Generation stops when this limit is reached.", + "group": "generation_length", + "type": "integer", + "default": 4096, + "range": { + "min": 1, + "max": 16384 + } + }, + { + "path": "frequency_penalty", + "label": "Frequency penalty", + "description": "Penalizes new tokens based on their existing frequency in the text so far, decreasing the model's likelihood to repeat the same line verbatim.", + "group": "sampling", + "type": "number", + "default": 0, + "range": { + "min": -2, + "max": 2 + } + }, + { + "path": "presence_penalty", + "label": "Presence penalty", + "description": "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.", + "group": "sampling", + "type": "number", + "default": 0, + "range": { + "min": -2, + "max": 2 + } + } + ] + }, { "provider": "nvidia", "authType": "api_key", @@ -35584,7 +35885,7 @@ const GENERATED_CATALOG = [ "type": "number", "default": 1, "range": { - "min": 0, + "min": 0.01, "max": 1, "step": 0.01 } diff --git a/packages/modelparams/src/generated/defaults.ts b/packages/modelparams/src/generated/defaults.ts index 2e9ba26c..d76420c6 100644 --- a/packages/modelparams/src/generated/defaults.ts +++ b/packages/modelparams/src/generated/defaults.ts @@ -57,6 +57,10 @@ export const DEFAULTS = { "extra_body.top_k": 20, "extra_body.enable_thinking": true, }, + "alibaba/pre-mock_new_gateway_session_affinity": { + "extra_body.top_k": 20, + "extra_body.enable_thinking": true, + }, "alibaba/qvq-max": { "extra_body.top_k": 20, "extra_body.enable_thinking": true, @@ -857,6 +861,7 @@ export const DEFAULTS = { "fireworks/minimax-m2p7": {}, "fireworks/minimax-m3": {}, "fireworks/muse-glimmer-30b": {}, + "fireworks/muse-spark-open": {}, "fireworks/nemotron-3-ultra-nvfp4": {}, "fireworks/nemotron-lightning-3p5-30b-a3b": {}, "fireworks/qwen3p7-plus": {}, @@ -1734,6 +1739,13 @@ export const DEFAULTS = { reasoning_effort: "high", reasoning_budget: 16384, }, + "nvidia/nemotron-3.5-super-vl-120b-a12b": { + temperature: 1, + top_p: 0.95, + max_tokens: 16384, + reasoning_effort: "high", + reasoning_budget: 16384, + }, "nvidia/nemotron-content-safety-reasoning-4b": { temperature: 1, top_p: 1, @@ -1753,6 +1765,13 @@ export const DEFAULTS = { frequency_penalty: 0, presence_penalty: 0, }, + "nvidia/step-5-preview": { + temperature: 0.6, + top_p: 0.95, + max_tokens: 4096, + frequency_penalty: 0, + presence_penalty: 0, + }, "nvidia/usdcode-llama-3.1-70b-instruct": { temperature: 0.1, top_p: 1, diff --git a/packages/modelparams/src/generated/model-ids.ts b/packages/modelparams/src/generated/model-ids.ts index f901e27e..9c8b5830 100644 --- a/packages/modelparams/src/generated/model-ids.ts +++ b/packages/modelparams/src/generated/model-ids.ts @@ -15,6 +15,7 @@ export const MODEL_IDS = [ "alibaba/glm-5.3-prime", "alibaba/kimi-k2.7-code", "alibaba/kimi-k3", + "alibaba/pre-mock_new_gateway_session_affinity", "alibaba/qvq-max", "alibaba/qwen-coder-plus", "alibaba/qwen-flash", @@ -214,6 +215,7 @@ export const MODEL_IDS = [ "fireworks/minimax-m2p7", "fireworks/minimax-m3", "fireworks/muse-glimmer-30b", + "fireworks/muse-spark-open", "fireworks/nemotron-3-ultra-nvfp4", "fireworks/nemotron-lightning-3p5-30b-a3b", "fireworks/qwen3p7-plus", @@ -352,9 +354,11 @@ export const MODEL_IDS = [ "nvidia/nemotron-3-ultra-550b-a55b", "nvidia/nemotron-3-ultra-subscription", "nvidia/nemotron-3.5-lightning-30b-a3b", + "nvidia/nemotron-3.5-super-vl-120b-a12b", "nvidia/nemotron-content-safety-reasoning-4b", "nvidia/nemotron-mini-4b-instruct", "nvidia/riva-translate-4b-instruct-v1.1", + "nvidia/step-5-preview", "nvidia/usdcode-llama-3.1-70b-instruct", "openai/chatgpt-4o-latest", "openai/gpt-3.5-turbo", diff --git a/packages/modelparams/src/generated/params-by-id.ts b/packages/modelparams/src/generated/params-by-id.ts index 0e3faaa9..681487d3 100644 --- a/packages/modelparams/src/generated/params-by-id.ts +++ b/packages/modelparams/src/generated/params-by-id.ts @@ -111,6 +111,14 @@ export type ParamsById = { "extra_body.enable_thinking": boolean; "extra_body.thinking_budget": number; }; + "alibaba/pre-mock_new_gateway_session_affinity": { + max_completion_tokens: number; + temperature: number; + top_p: number; + "extra_body.top_k": number; + "extra_body.enable_thinking": boolean; + "extra_body.thinking_budget": number; + }; "alibaba/qvq-max": { max_completion_tokens: number; temperature: number; @@ -1604,6 +1612,15 @@ export type ParamsById = { frequency_penalty: number; "response_format.type": "text" | "json_object"; }; + "fireworks/muse-spark-open": { + max_tokens: number; + temperature: number; + top_p: number; + top_k: number; + presence_penalty: number; + frequency_penalty: number; + "response_format.type": "text" | "json_object"; + }; "fireworks/nemotron-3-ultra-nvfp4": { max_tokens: number; temperature: number; @@ -2825,6 +2842,13 @@ export type ParamsById = { reasoning_effort: "none" | "low" | "high"; reasoning_budget: number; }; + "nvidia/nemotron-3.5-super-vl-120b-a12b": { + temperature: number; + top_p: number; + max_tokens: number; + reasoning_effort: "none" | "low" | "high"; + reasoning_budget: number; + }; "nvidia/nemotron-content-safety-reasoning-4b": { temperature: number; top_p: number; @@ -2848,6 +2872,13 @@ export type ParamsById = { presence_penalty: number; stop: string; }; + "nvidia/step-5-preview": { + temperature: number; + top_p: number; + max_tokens: number; + frequency_penalty: number; + presence_penalty: number; + }; "nvidia/usdcode-llama-3.1-70b-instruct": { temperature: number; top_p: number;