chore(catalog): sync from gateway — 1 added, 3 corrected, 0 retired

models/fireworks/inkling.yaml

@@ -26,7 +26,7 @@ params:
     label: Top P
     description: Controls nucleus sampling by limiting generation to tokens within the selected cumulative probability.
     range:
-      min: 0
+      min: 0.01
       max: 1
       step: 0.01
     group: sampling

models/google/gemini-3.7-flash.yaml

@@ -49,9 +49,8 @@ params:
     type: enum
     label: Thinking level
     description: Controls Gemini 3.1 Flash-Lite reasoning effort.
-    default: minimal
+    default: low
     values:
-      - minimal
       - low
       - medium
       - high

models/groq/gpt-oss-20b.yaml

@@ -77,7 +77,6 @@ params:
     description: Controls how reasoning tokens are returned — hidden from the response, raw within the content, or parsed into a separate field.
     values:
       - hidden
-      - raw
       - parsed
     group: reasoning
   - path: response_format.type

models/nvidia/llama-3.2-90b-vision-instruct.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: llama-3.2-90b-vision-instruct
+wireId: meta/llama-3.2-90b-vision-instruct
+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

packages/modelparams-python/src/modelparams/_generated/catalog.json

@@ -15542,7 +15542,7 @@
         "group": "sampling",
         "type": "number",
         "range": {
-          "min": 0,
+          "min": 0.01,
           "max": 1,
           "step": 0.01
         }
@@ -18393,9 +18393,8 @@
         "description": "Controls Gemini 3.1 Flash-Lite reasoning effort.",
         "group": "reasoning",
         "type": "enum",
-        "default": "minimal",
+        "default": "low",
         "values": [
-          "minimal",
           "low",
           "medium",
           "high"
@@ -19582,7 +19581,6 @@
         "type": "enum",
         "values": [
           "hidden",
-          "raw",
           "parsed"
         ]
       },
@@ -26648,6 +26646,74 @@
       }
     ]
   },
+  {
+    "provider": "nvidia",
+    "authType": "api_key",
+    "apiSurface": "openai-chat-completions",
+    "model": "llama-3.2-90b-vision-instruct",
+    "wireId": "meta/llama-3.2-90b-vision-instruct",
+    "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",

packages/modelparams-python/src/modelparams/_generated/model_ids.py

@@ -334,6 +334,7 @@
     "nvidia/llama-3.1-nemotron-safety-guard-8b-v3",
     "nvidia/llama-3.1-nemotron-ultra-253b-v1",
     "nvidia/llama-3.2-11b-vision-instruct",
+    "nvidia/llama-3.2-90b-vision-instruct",
     "nvidia/llama-3.3-nemotron-super-49b-v1",
     "nvidia/llama-3.3-nemotron-super-49b-v1.5",
     "nvidia/minimax-m3",
@@ -818,6 +819,7 @@
     "nvidia/llama-3.1-nemotron-safety-guard-8b-v3",
     "nvidia/llama-3.1-nemotron-ultra-253b-v1",
     "nvidia/llama-3.2-11b-vision-instruct",
+    "nvidia/llama-3.2-90b-vision-instruct",
     "nvidia/llama-3.3-nemotron-super-49b-v1",
     "nvidia/llama-3.3-nemotron-super-49b-v1.5",
     "nvidia/minimax-m3",

packages/modelparams-python/src/modelparams/_generated/registry.py

@@ -359,6 +359,7 @@
     "nvidia/llama-3.1-nemotron-safety-guard-8b-v3": nvidia.Llama_3_1_Nemotron_Safety_Guard_8b_V3Params,
     "nvidia/llama-3.1-nemotron-ultra-253b-v1": nvidia.Llama_3_1_Nemotron_Ultra_253b_V1Params,
     "nvidia/llama-3.2-11b-vision-instruct": nvidia.Llama_3_2_11b_Vision_InstructParams,
+    "nvidia/llama-3.2-90b-vision-instruct": nvidia.Llama_3_2_90b_Vision_InstructParams,
     "nvidia/llama-3.3-nemotron-super-49b-v1": nvidia.Llama_3_3_Nemotron_Super_49b_V1Params,
     "nvidia/llama-3.3-nemotron-super-49b-v1.5": nvidia.Llama_3_3_Nemotron_Super_49b_V1_5Params,
     "nvidia/minimax-m3": nvidia.Minimax_M3Params,

packages/modelparams-python/src/modelparams/types/fireworks.py

@@ -180,7 +180,7 @@
     {
         "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_p": Annotated[float, Field(ge=0.01, 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)],

packages/modelparams-python/src/modelparams/types/google.py

@@ -306,7 +306,7 @@
         "generationConfig.topP": Annotated[float, Field(ge=0, le=1)],
         "generationConfig.topK": Annotated[int, Field(ge=0)],
         "generationConfig.seed": int,
-        "generationConfig.thinkingConfig.thinkingLevel": Literal["minimal", "low", "medium", "high"],
+        "generationConfig.thinkingConfig.thinkingLevel": Literal["low", "medium", "high"],
         "generationConfig.thinkingConfig.includeThoughts": bool,
         "generationConfig.responseMimeType": Literal["text/plain", "application/json"],
     },

packages/modelparams-python/src/modelparams/types/groq.py

@@ -54,7 +54,7 @@
         "seed": int,
         "stop": str,
         "reasoning_effort": Literal["none", "default"],
-        "reasoning_format": Literal["hidden", "raw", "parsed"],
+        "reasoning_format": Literal["hidden", "parsed"],
         "response_format.type": Literal["text", "json_object"],
     },
     total=False,

packages/modelparams-python/src/modelparams/types/nvidia.py

@@ -221,6 +221,19 @@
 )
 setattr(Llama_3_2_11b_Vision_InstructParams, "__pydantic_config__", _PARAMS_CONFIG)
 
+Llama_3_2_90b_Vision_InstructParams = TypedDict(
+    "Llama_3_2_90b_Vision_InstructParams",
+    {
+        "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(Llama_3_2_90b_Vision_InstructParams, "__pydantic_config__", _PARAMS_CONFIG)
+
 Llama_3_3_Nemotron_Super_49b_V1Params = TypedDict(
     "Llama_3_3_Nemotron_Super_49b_V1Params",
     {
@@ -453,6 +466,7 @@
     "Llama_3_1_Nemotron_Safety_Guard_8b_V3Params",
     "Llama_3_1_Nemotron_Ultra_253b_V1Params",
     "Llama_3_2_11b_Vision_InstructParams",
+    "Llama_3_2_90b_Vision_InstructParams",
     "Llama_3_3_Nemotron_Super_49b_V1Params",
     "Llama_3_3_Nemotron_Super_49b_V1_5Params",
     "Minimax_M3Params",

packages/modelparams/src/generated/data.ts

@@ -15547,7 +15547,7 @@ const GENERATED_CATALOG = [
         "group": "sampling",
         "type": "number",
         "range": {
-          "min": 0,
+          "min": 0.01,
           "max": 1,
           "step": 0.01
         }
@@ -18398,9 +18398,8 @@ const GENERATED_CATALOG = [
         "description": "Controls Gemini 3.1 Flash-Lite reasoning effort.",
         "group": "reasoning",
         "type": "enum",
-        "default": "minimal",
+        "default": "low",
         "values": [
-          "minimal",
           "low",
           "medium",
           "high"
@@ -19587,7 +19586,6 @@ const GENERATED_CATALOG = [
         "type": "enum",
         "values": [
           "hidden",
-          "raw",
           "parsed"
         ]
       },
@@ -26653,6 +26651,74 @@ const GENERATED_CATALOG = [
       }
     ]
   },
+  {
+    "provider": "nvidia",
+    "authType": "api_key",
+    "apiSurface": "openai-chat-completions",
+    "model": "llama-3.2-90b-vision-instruct",
+    "wireId": "meta/llama-3.2-90b-vision-instruct",
+    "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",

packages/modelparams/src/generated/defaults.ts

@@ -990,7 +990,7 @@ export const DEFAULTS = {
     "generationConfig.temperature": 1,
     "generationConfig.topP": 0.95,
     "generationConfig.topK": 64,
-    "generationConfig.thinkingConfig.thinkingLevel": "minimal",
+    "generationConfig.thinkingConfig.thinkingLevel": "low",
     "generationConfig.thinkingConfig.includeThoughts": false,
     "generationConfig.responseMimeType": "text/plain",
   },
@@ -1611,6 +1611,13 @@ export const DEFAULTS = {
     frequency_penalty: 0,
     presence_penalty: 0,
   },
+  "nvidia/llama-3.2-90b-vision-instruct": {
+    temperature: 0.6,
+    top_p: 0.95,
+    max_tokens: 4096,
+    frequency_penalty: 0,
+    presence_penalty: 0,
+  },
   "nvidia/llama-3.3-nemotron-super-49b-v1": {
     temperature: 0.6,
     top_p: 0.95,

packages/modelparams/src/generated/model-ids.ts

@@ -332,6 +332,7 @@ export const MODEL_IDS = [
   "nvidia/llama-3.1-nemotron-safety-guard-8b-v3",
   "nvidia/llama-3.1-nemotron-ultra-253b-v1",
   "nvidia/llama-3.2-11b-vision-instruct",
+  "nvidia/llama-3.2-90b-vision-instruct",
   "nvidia/llama-3.3-nemotron-super-49b-v1",
   "nvidia/llama-3.3-nemotron-super-49b-v1.5",
   "nvidia/minimax-m3",