@@ -3598,6 +3598,231 @@ components:
35983598 oneOf:
35993599 - $ref: '#/components/schemas/ContainerAutoEnvironment'
36003600 - $ref: '#/components/schemas/ContainerReferenceEnvironment'
3601+ BenchmarkPricing:
3602+ description: OpenRouter pricing per token for this model. Null if pricing is unavailable.
3603+ example:
3604+ completion: '0.000015'
3605+ prompt: '0.000003'
3606+ nullable: true
3607+ properties:
3608+ completion:
3609+ description: Cost per output token (USD, decimal string).
3610+ example: '0.000015'
3611+ type: string
3612+ prompt:
3613+ description: Cost per input token (USD, decimal string).
3614+ example: '0.000003'
3615+ type: string
3616+ required:
3617+ - prompt
3618+ - completion
3619+ type: object
3620+ BenchmarksDAItem:
3621+ example:
3622+ arena: models
3623+ avg_generation_time_ms: 3200
3624+ category: codecategories
3625+ display_name: Claude Sonnet 4
3626+ elo: 1423
3627+ model_permaslug: anthropic/claude-sonnet-4
3628+ pricing:
3629+ completion: '0.000015'
3630+ prompt: '0.000003'
3631+ tournament_stats:
3632+ first_place: 12
3633+ fourth_place: 2
3634+ second_place: 8
3635+ third_place: 5
3636+ total: 27
3637+ win_rate: 72
3638+ properties:
3639+ arena:
3640+ description: Arena this ranking belongs to.
3641+ example: models
3642+ type: string
3643+ avg_generation_time_ms:
3644+ description: Average generation time in milliseconds.
3645+ example: 3200
3646+ format: double
3647+ nullable: true
3648+ type: number
3649+ category:
3650+ description: Category within the arena.
3651+ example: codecategories
3652+ type: string
3653+ display_name:
3654+ description: Human-readable model name from Design Arena.
3655+ example: Claude Sonnet 4
3656+ type: string
3657+ elo:
3658+ description: ELO rating from head-to-head arena battles.
3659+ example: 1423
3660+ format: double
3661+ type: number
3662+ model_permaslug:
3663+ description: >-
3664+ Stable OpenRouter model identifier when the model is on OpenRouter; otherwise the upstream Design Arena
3665+ model id. Use pricing != null to detect OpenRouter-mapped models.
3666+ example: anthropic/claude-sonnet-4
3667+ type: string
3668+ pricing:
3669+ $ref: '#/components/schemas/BenchmarkPricing'
3670+ tournament_stats:
3671+ description: Placement distribution from tournament matches.
3672+ properties:
3673+ first_place:
3674+ nullable: true
3675+ type: integer
3676+ fourth_place:
3677+ nullable: true
3678+ type: integer
3679+ second_place:
3680+ nullable: true
3681+ type: integer
3682+ third_place:
3683+ nullable: true
3684+ type: integer
3685+ total:
3686+ nullable: true
3687+ type: integer
3688+ required:
3689+ - first_place
3690+ - second_place
3691+ - third_place
3692+ - fourth_place
3693+ - total
3694+ type: object
3695+ win_rate:
3696+ description: Win rate as a percentage (0–100).
3697+ example: 72
3698+ format: double
3699+ type: number
3700+ required:
3701+ - model_permaslug
3702+ - display_name
3703+ - arena
3704+ - category
3705+ - elo
3706+ - win_rate
3707+ - avg_generation_time_ms
3708+ - tournament_stats
3709+ - pricing
3710+ type: object
3711+ BenchmarksDAMeta:
3712+ example:
3713+ arena: models
3714+ as_of: '2026-06-03T12:00:00Z'
3715+ category: null
3716+ citation: 'Source: Design Arena (www.designarena.ai) via OpenRouter (openrouter.ai/rankings).'
3717+ elo_bounds:
3718+ max: 1600
3719+ min: 900
3720+ model_count: 50
3721+ source: design-arena
3722+ source_url: https://www.designarena.ai
3723+ version: v1
3724+ properties:
3725+ arena:
3726+ description: The arena filter applied.
3727+ type: string
3728+ as_of:
3729+ description: ISO-8601 timestamp of when this data was generated.
3730+ example: '2026-06-03T12:00:00Z'
3731+ type: string
3732+ category:
3733+ description: The category filter applied, or null if showing all.
3734+ nullable: true
3735+ type: string
3736+ citation:
3737+ description: Required attribution when republishing this data.
3738+ example: 'Source: Design Arena (www.designarena.ai) via OpenRouter (openrouter.ai/rankings).'
3739+ type: string
3740+ elo_bounds:
3741+ description: ELO range across all returned models for normalization.
3742+ properties:
3743+ max:
3744+ description: Maximum ELO in the result set.
3745+ format: double
3746+ type: number
3747+ min:
3748+ description: Minimum ELO in the result set.
3749+ format: double
3750+ type: number
3751+ required:
3752+ - min
3753+ - max
3754+ type: object
3755+ model_count:
3756+ description: Number of unique models in the response.
3757+ type: integer
3758+ source:
3759+ description: Data source identifier.
3760+ enum:
3761+ - design-arena
3762+ type: string
3763+ source_url:
3764+ description: URL of the upstream data source.
3765+ enum:
3766+ - https://www.designarena.ai
3767+ type: string
3768+ version:
3769+ description: Dataset version.
3770+ enum:
3771+ - v1
3772+ type: string
3773+ required:
3774+ - as_of
3775+ - version
3776+ - source
3777+ - source_url
3778+ - citation
3779+ - model_count
3780+ - arena
3781+ - category
3782+ - elo_bounds
3783+ type: object
3784+ BenchmarksDAResponse:
3785+ example:
3786+ data:
3787+ - arena: models
3788+ avg_generation_time_ms: 3200
3789+ category: codecategories
3790+ display_name: Claude Sonnet 4
3791+ elo: 1423
3792+ model_permaslug: anthropic/claude-sonnet-4
3793+ pricing:
3794+ completion: '0.000015'
3795+ prompt: '0.000003'
3796+ tournament_stats:
3797+ first_place: 12
3798+ fourth_place: 2
3799+ second_place: 8
3800+ third_place: 5
3801+ total: 27
3802+ win_rate: 72
3803+ meta:
3804+ arena: models
3805+ as_of: '2026-06-03T12:00:00Z'
3806+ category: null
3807+ citation: 'Source: Design Arena (www.designarena.ai) via OpenRouter (openrouter.ai/rankings).'
3808+ elo_bounds:
3809+ max: 1600
3810+ min: 900
3811+ model_count: 1
3812+ source: design-arena
3813+ source_url: https://www.designarena.ai
3814+ version: v1
3815+ properties:
3816+ data:
3817+ items:
3818+ $ref: '#/components/schemas/BenchmarksDAItem'
3819+ type: array
3820+ meta:
3821+ $ref: '#/components/schemas/BenchmarksDAMeta'
3822+ required:
3823+ - data
3824+ - meta
3825+ type: object
36013826 BigNumberUnion:
36023827 description: Price per million prompt tokens
36033828 example: 1000
@@ -23316,6 +23541,130 @@ paths:
2331623541 outputs:
2331723542 results: $.data
2331823543 type: offsetLimit
23544+ /datasets/benchmarks/design-arena:
23545+ get:
23546+ description: >-
23547+ Returns ELO ratings from head-to-head arena battles on Design Arena. Filterable by arena
23548+ (models/builders/agents) and category. Includes OpenRouter pricing per model. Authenticate with any valid
23549+ OpenRouter API key. Rate-limited to 30 requests/minute per key and 500 requests/day per account.
23550+ operationId: getBenchmarksDesignArena
23551+ parameters:
23552+ - description: Arena to query. Defaults to `models`.
23553+ in: query
23554+ name: arena
23555+ required: false
23556+ schema:
23557+ default: models
23558+ description: Arena to query. Defaults to `models`.
23559+ enum:
23560+ - models
23561+ - builders
23562+ - agents
23563+ example: models
23564+ type: string
23565+ - description: >-
23566+ Category within the arena (e.g. `codecategories`, `uicomponent`, `gamedev`, `3d`, `dataviz`, `image`,
23567+ `video`, `svg`). When omitted, returns all categories.
23568+ in: query
23569+ name: category
23570+ required: false
23571+ schema:
23572+ description: >-
23573+ Category within the arena (e.g. `codecategories`, `uicomponent`, `gamedev`, `3d`, `dataviz`, `image`,
23574+ `video`, `svg`). When omitted, returns all categories.
23575+ example: codecategories
23576+ type: string
23577+ - description: 'Max results to return: per category when no category filter is applied (1–100, default 50).'
23578+ in: query
23579+ name: max_results
23580+ required: false
23581+ schema:
23582+ default: 50
23583+ description: 'Max results to return: per category when no category filter is applied (1–100, default 50).'
23584+ example: 20
23585+ maximum: 100
23586+ minimum: 1
23587+ type: integer
23588+ responses:
23589+ '200':
23590+ content:
23591+ application/json:
23592+ example:
23593+ data:
23594+ - arena: models
23595+ avg_generation_time_ms: 3200
23596+ category: codecategories
23597+ display_name: Claude Sonnet 4
23598+ elo: 1423
23599+ model_permaslug: anthropic/claude-sonnet-4
23600+ pricing:
23601+ completion: '0.000015'
23602+ prompt: '0.000003'
23603+ tournament_stats:
23604+ first_place: 12
23605+ fourth_place: 2
23606+ second_place: 8
23607+ third_place: 5
23608+ total: 27
23609+ win_rate: 72
23610+ meta:
23611+ arena: models
23612+ as_of: '2026-06-03T12:00:00Z'
23613+ category: null
23614+ citation: 'Source: Design Arena (www.designarena.ai) via OpenRouter (openrouter.ai/rankings).'
23615+ elo_bounds:
23616+ max: 1600
23617+ min: 900
23618+ model_count: 1
23619+ source: design-arena
23620+ source_url: https://www.designarena.ai
23621+ version: v1
23622+ schema:
23623+ $ref: '#/components/schemas/BenchmarksDAResponse'
23624+ description: Design Arena ELO rankings with pricing and attribution metadata.
23625+ '400':
23626+ content:
23627+ application/json:
23628+ example:
23629+ error:
23630+ code: 400
23631+ message: Invalid request parameters
23632+ schema:
23633+ $ref: '#/components/schemas/BadRequestResponse'
23634+ description: Bad Request - Invalid request parameters or malformed input
23635+ '401':
23636+ content:
23637+ application/json:
23638+ example:
23639+ error:
23640+ code: 401
23641+ message: Missing Authentication header
23642+ schema:
23643+ $ref: '#/components/schemas/UnauthorizedResponse'
23644+ description: Unauthorized - Authentication required or invalid credentials
23645+ '429':
23646+ content:
23647+ application/json:
23648+ example:
23649+ error:
23650+ code: 429
23651+ message: Rate limit exceeded
23652+ schema:
23653+ $ref: '#/components/schemas/TooManyRequestsResponse'
23654+ description: Too Many Requests - Rate limit exceeded
23655+ '500':
23656+ content:
23657+ application/json:
23658+ example:
23659+ error:
23660+ code: 500
23661+ message: Internal Server Error
23662+ schema:
23663+ $ref: '#/components/schemas/InternalServerResponse'
23664+ description: Internal Server Error - Unexpected server error
23665+ summary: Design Arena Benchmark Rankings
23666+ tags:
23667+ - Datasets
2331923668 /datasets/rankings-daily:
2332023669 get:
2332123670 description: |-
0 commit comments