@@ -12202,6 +12202,15 @@ components:
1220212202 image: '0'
1220312203 prompt: '0.00003'
1220412204 request: '0'
12205+ reasoning:
12206+ default_effort: medium
12207+ default_enabled: true
12208+ mandatory: false
12209+ supported_efforts:
12210+ - high
12211+ - medium
12212+ - low
12213+ - minimal
1220512214 supported_parameters:
1220612215 - temperature
1220712216 - top_p
@@ -12264,6 +12273,8 @@ components:
1226412273 $ref: '#/components/schemas/PerRequestLimits'
1226512274 pricing:
1226612275 $ref: '#/components/schemas/PublicPricing'
12276+ reasoning:
12277+ $ref: '#/components/schemas/ModelReasoning'
1226712278 supported_parameters:
1226812279 description: List of supported parameters for this model
1226912280 items:
@@ -12426,6 +12437,47 @@ components:
1242612437 description: Model to use for completion
1242712438 example: openai/gpt-4
1242812439 type: string
12440+ ModelReasoning:
12441+ description: Reasoning effort configuration. Omitted for non-reasoning models and dynamic router models.
12442+ example:
12443+ default_effort: medium
12444+ default_enabled: true
12445+ mandatory: false
12446+ supported_efforts:
12447+ - high
12448+ - medium
12449+ - low
12450+ - minimal
12451+ properties:
12452+ default_effort:
12453+ allOf:
12454+ - $ref: '#/components/schemas/ReasoningEffort'
12455+ - description: >-
12456+ Default reasoning effort when the client enables reasoning without specifying effort. Maps to
12457+ `reasoning.effort` in chat requests. When `"none"`, prefer omitting effort unless the user explicitly
12458+ disables reasoning.
12459+ default_enabled:
12460+ description: Default reasoning enabled state when the client does not set `reasoning.enabled`.
12461+ type: boolean
12462+ mandatory:
12463+ description: When true, reasoning cannot be disabled and effort "none" is rejected.
12464+ type: boolean
12465+ supported_efforts:
12466+ description: >-
12467+ Allowed reasoning effort values for this model, in descending effort order (highest first). Null means no
12468+ allowlist — all gateway effort values are accepted.
12469+ items:
12470+ $ref: '#/components/schemas/ReasoningEffort'
12471+ nullable: true
12472+ type: array
12473+ supports_max_tokens:
12474+ description: >-
12475+ Present and `true` when the model accepts `reasoning.max_tokens` in requests (Anthropic-style) instead of or
12476+ in addition to `reasoning.effort`. Omitted otherwise.
12477+ type: boolean
12478+ required:
12479+ - mandatory
12480+ type: object
1242912481 ModelResponse:
1243012482 description: Single model response
1243112483 example:
@@ -28837,7 +28889,7 @@ paths:
2883728889 id: openai/gpt-4
2883828890 knowledge_cutoff: null
2883928891 links:
28840- details: /api/v1/models/openai/gpt-5. 4/endpoints
28892+ details: /api/v1/models/openai/gpt-4/endpoints
2884128893 name: GPT-4
2884228894 per_request_limits: null
2884328895 pricing:
0 commit comments