[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill-openrouter-openrouter-benchmarks":3,"mdc-jeyou1-key":33,"related-repo-openrouter-openrouter-benchmarks":1038,"related-org-openrouter-openrouter-benchmarks":1125},{"slug":4,"name":4,"fn":5,"description":6,"org":7,"tags":12,"stars":23,"repoUrl":24,"updatedAt":25,"license":26,"forks":27,"topics":28,"repo":29,"sourceUrl":31,"mdContent":32},"openrouter-benchmarks","query OpenRouter model benchmarks","Query OpenRouter's Benchmarks API for model benchmark rankings and scores. Use when the user asks for benchmark-backed model selection, model rankings by coding\u002Fintelligence\u002Fagentic ability, Artificial Analysis or Design Arena ELO\u002Fwin-rate results, benchmark citations, or wants to call GET \u002Fapi\u002Fv1\u002Fbenchmarks. Also use alongside openrouter-models when the user asks what model should power an app, product, workflow, or use case and benchmark evidence could inform or rule out part of the recommendation, including creative writing, editing, coding, design, agentic, or intelligence-heavy apps. Do not use for OpenRouter usage analytics, billing\u002Fspend analysis, generation metadata, provider uptime\u002Flatency, generic model pricing\u002Fcapability lookup without any selection or benchmark-relevance decision, or creating an evaluation suite for a local app.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":11},"openrouter","OpenRouter","https:\u002F\u002Fpexgzepcugksgbtrxkhf.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Forg-logos\u002Fopenrouter.png","OpenRouterTeam",[13,17,20],{"name":14,"slug":15,"type":16},"Benchmarking","benchmarking","tag",{"name":18,"slug":19,"type":16},"LLM","llm",{"name":21,"slug":22,"type":16},"Analytics","analytics",187,"https:\u002F\u002Fgithub.com\u002FOpenRouterTeam\u002Fskills","2026-07-14T05:38:27.658475",null,26,[],{"repoUrl":24,"stars":23,"forks":27,"topics":30,"description":26},[],"https:\u002F\u002Fgithub.com\u002FOpenRouterTeam\u002Fskills\u002Ftree\u002FHEAD\u002Fskills\u002Fopenrouter-benchmarks","---\nname: openrouter-benchmarks\ndescription: Query OpenRouter's Benchmarks API for model benchmark rankings and scores. Use when the user asks for benchmark-backed model selection, model rankings by coding\u002Fintelligence\u002Fagentic ability, Artificial Analysis or Design Arena ELO\u002Fwin-rate results, benchmark citations, or wants to call GET \u002Fapi\u002Fv1\u002Fbenchmarks. Also use alongside openrouter-models when the user asks what model should power an app, product, workflow, or use case and benchmark evidence could inform or rule out part of the recommendation, including creative writing, editing, coding, design, agentic, or intelligence-heavy apps. Do not use for OpenRouter usage analytics, billing\u002Fspend analysis, generation metadata, provider uptime\u002Flatency, generic model pricing\u002Fcapability lookup without any selection or benchmark-relevance decision, or creating an evaluation suite for a local app.\n---\n\n# OpenRouter Benchmarks\n\nUse OpenRouter's unified benchmarks endpoint to answer benchmark-backed model ranking and model-selection questions. The endpoint aggregates Artificial Analysis and Design Arena data and returns citation metadata that should be preserved when reporting results.\n\n## Prerequisites\n\nSet `OPENROUTER_API_KEY` to any valid OpenRouter API key. Benchmarks do not require a management key.\n\n```bash\nexport OPENROUTER_API_KEY=sk-or-v1-...\n```\n\n## Decision Tree\n\n| User wants to... | Action |\n|---|---|\n| See benchmark-ranked models across sources | Call `GET \u002Fapi\u002Fv1\u002Fbenchmarks` and preserve source\u002Fcitation metadata |\n| Choose a model for an app\u002Fuse case | Check whether Artificial Analysis or Design Arena contains a relevant signal; say when no direct benchmark exists |\n| Find best coding, intelligence, or agentic models | Use `task_type=coding`, `task_type=intelligence`, or `task_type=agentic` |\n| Query Artificial Analysis only | Use `source=artificial-analysis` |\n| Query Design Arena only | Use `source=design-arena`, plus `arena` and `category` when relevant |\n| Get raw API-shaped data for integration work | Return the raw `data`\u002F`meta` shape from the endpoint |\n| Understand all response fields or direct curl usage | Read `references\u002Fbenchmarks-api.md` |\n\nUse `openrouter-models` instead when the user needs pricing, context length, supported parameters, modalities, or provider endpoint performance without asking for benchmark rankings.\n\nFor creative writing, storytelling, or editorial apps, this endpoint currently has no direct writing-quality benchmark. Treat Artificial Analysis `intelligence_index` as a weak general-capability signal, and use `agentic_index` only if the app performs multi-step planning\u002Frevision. Do not imply that Design Arena visual\u002Fcode categories measure prose quality.\n\n## Availability Gate\n\nDo not recommend a benchmark-ranked model until it passes an availability check through the models\u002Fendpoints API. Benchmark rows can contain dated or benchmark-specific `model_permaslug` values that are useful for attribution but are not always the exact routable OpenRouter model ID.\n\nBefore recommending a benchmark candidate:\n\n1. Check `GET \u002Fapi\u002Fv1\u002Fmodels` for an exact `id` match to the benchmark `model_permaslug`.\n2. If there is no exact `id` match but a model has `canonical_slug` equal to the benchmark `model_permaslug`, treat the benchmark row as evidence for that model family, not as a directly recommendable ID. Use the model's actual `id` only after verifying availability.\n3. Check `GET \u002Fapi\u002Fv1\u002Fmodels\u002F{author}\u002F{slug}\u002Fendpoints` or use `openrouter-models` `get-endpoints.ts` for provider status.\n4. Prefer candidates with at least one clearly usable endpoint. If all endpoints are degraded, have `uptime_last_30m: 0`, or the OpenRouter model page\u002FAPI indicates the model is unavailable, exclude it from primary recommendations and explain that the benchmark result is not currently actionable.\n5. When availability is ambiguous, say so and recommend a verified available alternative instead of presenting the benchmark leader as the default choice.\n\nDo not rely on endpoint `status: 0` alone. Model-level availability signals such as routing error messages, warning messages, zero request limits, empty endpoint lists, or provider-specific access restrictions can make a benchmark leader non-actionable even when one endpoint appears operational. If availability signals disagree, explain the ambiguity and avoid making that model the primary recommendation.\n\n## API Usage\n\nQuery parameters:\n\n| Flag | Values | Notes |\n|---|---|---|\n| `source` | `artificial-analysis`, `design-arena` | Omitting it returns all sources. |\n| `task_type` | `coding`, `intelligence`, `agentic` | Maps to source-specific indices\u002Fcategories. |\n| `arena` | `models`, `builders`, `agents` | Design Arena only; defaults server-side to `models`. |\n| `category` | `codecategories`, `uicomponent`, `gamedev`, `3d`, `dataviz`, `image`, `video`, `svg`, etc. | Design Arena only. |\n| `max_results` | positive integer | Maximum number of rows returned by the API. |\n\nAlways preserve `meta.citation`, `meta.source_url`, and `meta.as_of`; include attribution when republishing benchmark data.\n\nWhen results include both sources, do not present them as a single absolute leaderboard: Artificial Analysis indices and Design Arena ELO use different scales. Compare within each source, or rerun with `source=artificial-analysis` or `source=design-arena` for a source-specific ranking.\n\n## Interpreting Results\n\n- Artificial Analysis rows include `intelligence_index`, `coding_index`, and `agentic_index`; higher is better.\n- Design Arena rows include `elo`, `win_rate`, `avg_generation_time_ms`, `arena`, `category`, and `tournament_stats`; higher `elo`\u002F`win_rate` is better, lower generation time is faster.\n- `pricing.prompt` and `pricing.completion` are USD per token as decimal strings. Multiply by 1,000,000 for per-million-token costs.\n- `model_permaslug` identifies the benchmarked model entry. Verify it against `GET \u002Fapi\u002Fv1\u002Fmodels` before using it as a chat\u002Fcompletions model ID.\n- `meta.model_count` counts unique models in the response, which can differ from `data.length` when multiple Design Arena categories are returned.\n\n## Direct API Call\n\n```bash\ncurl 'https:\u002F\u002Fopenrouter.ai\u002Fapi\u002Fv1\u002Fbenchmarks?source=artificial-analysis&task_type=coding&max_results=10' \\\n  -H \"Authorization: Bearer $OPENROUTER_API_KEY\"\n```\n\nRead `references\u002Fbenchmarks-api.md` when implementing against the raw API or handling source-specific response shapes.\n",{"data":34,"body":35},{"name":4,"description":6},{"type":36,"children":37},"root",[38,46,52,59,73,113,119,318,330,351,357,370,375,487,500,506,511,754,782,801,807,949,955,1021,1032],{"type":39,"tag":40,"props":41,"children":42},"element","h1",{"id":4},[43],{"type":44,"value":45},"text","OpenRouter Benchmarks",{"type":39,"tag":47,"props":48,"children":49},"p",{},[50],{"type":44,"value":51},"Use OpenRouter's unified benchmarks endpoint to answer benchmark-backed model ranking and model-selection questions. The endpoint aggregates Artificial Analysis and Design Arena data and returns citation metadata that should be preserved when reporting results.",{"type":39,"tag":53,"props":54,"children":56},"h2",{"id":55},"prerequisites",[57],{"type":44,"value":58},"Prerequisites",{"type":39,"tag":47,"props":60,"children":61},{},[62,64,71],{"type":44,"value":63},"Set ",{"type":39,"tag":65,"props":66,"children":68},"code",{"className":67},[],[69],{"type":44,"value":70},"OPENROUTER_API_KEY",{"type":44,"value":72}," to any valid OpenRouter API key. Benchmarks do not require a management key.",{"type":39,"tag":74,"props":75,"children":80},"pre",{"className":76,"code":77,"language":78,"meta":79,"style":79},"language-bash shiki shiki-themes material-theme-lighter material-theme material-theme-palenight","export OPENROUTER_API_KEY=sk-or-v1-...\n","bash","",[81],{"type":39,"tag":65,"props":82,"children":83},{"__ignoreMap":79},[84],{"type":39,"tag":85,"props":86,"children":89},"span",{"class":87,"line":88},"line",1,[90,96,102,108],{"type":39,"tag":85,"props":91,"children":93},{"style":92},"--shiki-light:#9C3EDA;--shiki-default:#C792EA;--shiki-dark:#C792EA",[94],{"type":44,"value":95},"export",{"type":39,"tag":85,"props":97,"children":99},{"style":98},"--shiki-light:#90A4AE;--shiki-default:#EEFFFF;--shiki-dark:#BABED8",[100],{"type":44,"value":101}," OPENROUTER_API_KEY",{"type":39,"tag":85,"props":103,"children":105},{"style":104},"--shiki-light:#39ADB5;--shiki-default:#89DDFF;--shiki-dark:#89DDFF",[106],{"type":44,"value":107},"=",{"type":39,"tag":85,"props":109,"children":110},{"style":98},[111],{"type":44,"value":112},"sk-or-v1-...\n",{"type":39,"tag":53,"props":114,"children":116},{"id":115},"decision-tree",[117],{"type":44,"value":118},"Decision Tree",{"type":39,"tag":120,"props":121,"children":122},"table",{},[123,142],{"type":39,"tag":124,"props":125,"children":126},"thead",{},[127],{"type":39,"tag":128,"props":129,"children":130},"tr",{},[131,137],{"type":39,"tag":132,"props":133,"children":134},"th",{},[135],{"type":44,"value":136},"User wants to...",{"type":39,"tag":132,"props":138,"children":139},{},[140],{"type":44,"value":141},"Action",{"type":39,"tag":143,"props":144,"children":145},"tbody",{},[146,168,181,216,234,270,299],{"type":39,"tag":128,"props":147,"children":148},{},[149,155],{"type":39,"tag":150,"props":151,"children":152},"td",{},[153],{"type":44,"value":154},"See benchmark-ranked models across sources",{"type":39,"tag":150,"props":156,"children":157},{},[158,160,166],{"type":44,"value":159},"Call ",{"type":39,"tag":65,"props":161,"children":163},{"className":162},[],[164],{"type":44,"value":165},"GET \u002Fapi\u002Fv1\u002Fbenchmarks",{"type":44,"value":167}," and preserve source\u002Fcitation metadata",{"type":39,"tag":128,"props":169,"children":170},{},[171,176],{"type":39,"tag":150,"props":172,"children":173},{},[174],{"type":44,"value":175},"Choose a model for an app\u002Fuse case",{"type":39,"tag":150,"props":177,"children":178},{},[179],{"type":44,"value":180},"Check whether Artificial Analysis or Design Arena contains a relevant signal; say when no direct benchmark exists",{"type":39,"tag":128,"props":182,"children":183},{},[184,189],{"type":39,"tag":150,"props":185,"children":186},{},[187],{"type":44,"value":188},"Find best coding, intelligence, or agentic models",{"type":39,"tag":150,"props":190,"children":191},{},[192,194,200,202,208,210],{"type":44,"value":193},"Use ",{"type":39,"tag":65,"props":195,"children":197},{"className":196},[],[198],{"type":44,"value":199},"task_type=coding",{"type":44,"value":201},", ",{"type":39,"tag":65,"props":203,"children":205},{"className":204},[],[206],{"type":44,"value":207},"task_type=intelligence",{"type":44,"value":209},", or ",{"type":39,"tag":65,"props":211,"children":213},{"className":212},[],[214],{"type":44,"value":215},"task_type=agentic",{"type":39,"tag":128,"props":217,"children":218},{},[219,224],{"type":39,"tag":150,"props":220,"children":221},{},[222],{"type":44,"value":223},"Query Artificial Analysis only",{"type":39,"tag":150,"props":225,"children":226},{},[227,228],{"type":44,"value":193},{"type":39,"tag":65,"props":229,"children":231},{"className":230},[],[232],{"type":44,"value":233},"source=artificial-analysis",{"type":39,"tag":128,"props":235,"children":236},{},[237,242],{"type":39,"tag":150,"props":238,"children":239},{},[240],{"type":44,"value":241},"Query Design Arena only",{"type":39,"tag":150,"props":243,"children":244},{},[245,246,252,254,260,262,268],{"type":44,"value":193},{"type":39,"tag":65,"props":247,"children":249},{"className":248},[],[250],{"type":44,"value":251},"source=design-arena",{"type":44,"value":253},", plus ",{"type":39,"tag":65,"props":255,"children":257},{"className":256},[],[258],{"type":44,"value":259},"arena",{"type":44,"value":261}," and ",{"type":39,"tag":65,"props":263,"children":265},{"className":264},[],[266],{"type":44,"value":267},"category",{"type":44,"value":269}," when relevant",{"type":39,"tag":128,"props":271,"children":272},{},[273,278],{"type":39,"tag":150,"props":274,"children":275},{},[276],{"type":44,"value":277},"Get raw API-shaped data for integration work",{"type":39,"tag":150,"props":279,"children":280},{},[281,283,289,291,297],{"type":44,"value":282},"Return the raw ",{"type":39,"tag":65,"props":284,"children":286},{"className":285},[],[287],{"type":44,"value":288},"data",{"type":44,"value":290},"\u002F",{"type":39,"tag":65,"props":292,"children":294},{"className":293},[],[295],{"type":44,"value":296},"meta",{"type":44,"value":298}," shape from the endpoint",{"type":39,"tag":128,"props":300,"children":301},{},[302,307],{"type":39,"tag":150,"props":303,"children":304},{},[305],{"type":44,"value":306},"Understand all response fields or direct curl usage",{"type":39,"tag":150,"props":308,"children":309},{},[310,312],{"type":44,"value":311},"Read ",{"type":39,"tag":65,"props":313,"children":315},{"className":314},[],[316],{"type":44,"value":317},"references\u002Fbenchmarks-api.md",{"type":39,"tag":47,"props":319,"children":320},{},[321,322,328],{"type":44,"value":193},{"type":39,"tag":65,"props":323,"children":325},{"className":324},[],[326],{"type":44,"value":327},"openrouter-models",{"type":44,"value":329}," instead when the user needs pricing, context length, supported parameters, modalities, or provider endpoint performance without asking for benchmark rankings.",{"type":39,"tag":47,"props":331,"children":332},{},[333,335,341,343,349],{"type":44,"value":334},"For creative writing, storytelling, or editorial apps, this endpoint currently has no direct writing-quality benchmark. Treat Artificial Analysis ",{"type":39,"tag":65,"props":336,"children":338},{"className":337},[],[339],{"type":44,"value":340},"intelligence_index",{"type":44,"value":342}," as a weak general-capability signal, and use ",{"type":39,"tag":65,"props":344,"children":346},{"className":345},[],[347],{"type":44,"value":348},"agentic_index",{"type":44,"value":350}," only if the app performs multi-step planning\u002Frevision. Do not imply that Design Arena visual\u002Fcode categories measure prose quality.",{"type":39,"tag":53,"props":352,"children":354},{"id":353},"availability-gate",[355],{"type":44,"value":356},"Availability Gate",{"type":39,"tag":47,"props":358,"children":359},{},[360,362,368],{"type":44,"value":361},"Do not recommend a benchmark-ranked model until it passes an availability check through the models\u002Fendpoints API. Benchmark rows can contain dated or benchmark-specific ",{"type":39,"tag":65,"props":363,"children":365},{"className":364},[],[366],{"type":44,"value":367},"model_permaslug",{"type":44,"value":369}," values that are useful for attribution but are not always the exact routable OpenRouter model ID.",{"type":39,"tag":47,"props":371,"children":372},{},[373],{"type":44,"value":374},"Before recommending a benchmark candidate:",{"type":39,"tag":376,"props":377,"children":378},"ol",{},[379,408,442,469,482],{"type":39,"tag":380,"props":381,"children":382},"li",{},[383,385,391,393,399,401,406],{"type":44,"value":384},"Check ",{"type":39,"tag":65,"props":386,"children":388},{"className":387},[],[389],{"type":44,"value":390},"GET \u002Fapi\u002Fv1\u002Fmodels",{"type":44,"value":392}," for an exact ",{"type":39,"tag":65,"props":394,"children":396},{"className":395},[],[397],{"type":44,"value":398},"id",{"type":44,"value":400}," match to the benchmark ",{"type":39,"tag":65,"props":402,"children":404},{"className":403},[],[405],{"type":44,"value":367},{"type":44,"value":407},".",{"type":39,"tag":380,"props":409,"children":410},{},[411,413,418,420,426,428,433,435,440],{"type":44,"value":412},"If there is no exact ",{"type":39,"tag":65,"props":414,"children":416},{"className":415},[],[417],{"type":44,"value":398},{"type":44,"value":419}," match but a model has ",{"type":39,"tag":65,"props":421,"children":423},{"className":422},[],[424],{"type":44,"value":425},"canonical_slug",{"type":44,"value":427}," equal to the benchmark ",{"type":39,"tag":65,"props":429,"children":431},{"className":430},[],[432],{"type":44,"value":367},{"type":44,"value":434},", treat the benchmark row as evidence for that model family, not as a directly recommendable ID. Use the model's actual ",{"type":39,"tag":65,"props":436,"children":438},{"className":437},[],[439],{"type":44,"value":398},{"type":44,"value":441}," only after verifying availability.",{"type":39,"tag":380,"props":443,"children":444},{},[445,446,452,454,459,461,467],{"type":44,"value":384},{"type":39,"tag":65,"props":447,"children":449},{"className":448},[],[450],{"type":44,"value":451},"GET \u002Fapi\u002Fv1\u002Fmodels\u002F{author}\u002F{slug}\u002Fendpoints",{"type":44,"value":453}," or use ",{"type":39,"tag":65,"props":455,"children":457},{"className":456},[],[458],{"type":44,"value":327},{"type":44,"value":460}," ",{"type":39,"tag":65,"props":462,"children":464},{"className":463},[],[465],{"type":44,"value":466},"get-endpoints.ts",{"type":44,"value":468}," for provider status.",{"type":39,"tag":380,"props":470,"children":471},{},[472,474,480],{"type":44,"value":473},"Prefer candidates with at least one clearly usable endpoint. If all endpoints are degraded, have ",{"type":39,"tag":65,"props":475,"children":477},{"className":476},[],[478],{"type":44,"value":479},"uptime_last_30m: 0",{"type":44,"value":481},", or the OpenRouter model page\u002FAPI indicates the model is unavailable, exclude it from primary recommendations and explain that the benchmark result is not currently actionable.",{"type":39,"tag":380,"props":483,"children":484},{},[485],{"type":44,"value":486},"When availability is ambiguous, say so and recommend a verified available alternative instead of presenting the benchmark leader as the default choice.",{"type":39,"tag":47,"props":488,"children":489},{},[490,492,498],{"type":44,"value":491},"Do not rely on endpoint ",{"type":39,"tag":65,"props":493,"children":495},{"className":494},[],[496],{"type":44,"value":497},"status: 0",{"type":44,"value":499}," alone. Model-level availability signals such as routing error messages, warning messages, zero request limits, empty endpoint lists, or provider-specific access restrictions can make a benchmark leader non-actionable even when one endpoint appears operational. If availability signals disagree, explain the ambiguity and avoid making that model the primary recommendation.",{"type":39,"tag":53,"props":501,"children":503},{"id":502},"api-usage",[504],{"type":44,"value":505},"API Usage",{"type":39,"tag":47,"props":507,"children":508},{},[509],{"type":44,"value":510},"Query parameters:",{"type":39,"tag":120,"props":512,"children":513},{},[514,535],{"type":39,"tag":124,"props":515,"children":516},{},[517],{"type":39,"tag":128,"props":518,"children":519},{},[520,525,530],{"type":39,"tag":132,"props":521,"children":522},{},[523],{"type":44,"value":524},"Flag",{"type":39,"tag":132,"props":526,"children":527},{},[528],{"type":44,"value":529},"Values",{"type":39,"tag":132,"props":531,"children":532},{},[533],{"type":44,"value":534},"Notes",{"type":39,"tag":143,"props":536,"children":537},{},[538,571,611,656,732],{"type":39,"tag":128,"props":539,"children":540},{},[541,550,566],{"type":39,"tag":150,"props":542,"children":543},{},[544],{"type":39,"tag":65,"props":545,"children":547},{"className":546},[],[548],{"type":44,"value":549},"source",{"type":39,"tag":150,"props":551,"children":552},{},[553,559,560],{"type":39,"tag":65,"props":554,"children":556},{"className":555},[],[557],{"type":44,"value":558},"artificial-analysis",{"type":44,"value":201},{"type":39,"tag":65,"props":561,"children":563},{"className":562},[],[564],{"type":44,"value":565},"design-arena",{"type":39,"tag":150,"props":567,"children":568},{},[569],{"type":44,"value":570},"Omitting it returns all sources.",{"type":39,"tag":128,"props":572,"children":573},{},[574,583,606],{"type":39,"tag":150,"props":575,"children":576},{},[577],{"type":39,"tag":65,"props":578,"children":580},{"className":579},[],[581],{"type":44,"value":582},"task_type",{"type":39,"tag":150,"props":584,"children":585},{},[586,592,593,599,600],{"type":39,"tag":65,"props":587,"children":589},{"className":588},[],[590],{"type":44,"value":591},"coding",{"type":44,"value":201},{"type":39,"tag":65,"props":594,"children":596},{"className":595},[],[597],{"type":44,"value":598},"intelligence",{"type":44,"value":201},{"type":39,"tag":65,"props":601,"children":603},{"className":602},[],[604],{"type":44,"value":605},"agentic",{"type":39,"tag":150,"props":607,"children":608},{},[609],{"type":44,"value":610},"Maps to source-specific indices\u002Fcategories.",{"type":39,"tag":128,"props":612,"children":613},{},[614,622,645],{"type":39,"tag":150,"props":615,"children":616},{},[617],{"type":39,"tag":65,"props":618,"children":620},{"className":619},[],[621],{"type":44,"value":259},{"type":39,"tag":150,"props":623,"children":624},{},[625,631,632,638,639],{"type":39,"tag":65,"props":626,"children":628},{"className":627},[],[629],{"type":44,"value":630},"models",{"type":44,"value":201},{"type":39,"tag":65,"props":633,"children":635},{"className":634},[],[636],{"type":44,"value":637},"builders",{"type":44,"value":201},{"type":39,"tag":65,"props":640,"children":642},{"className":641},[],[643],{"type":44,"value":644},"agents",{"type":39,"tag":150,"props":646,"children":647},{},[648,650,655],{"type":44,"value":649},"Design Arena only; defaults server-side to ",{"type":39,"tag":65,"props":651,"children":653},{"className":652},[],[654],{"type":44,"value":630},{"type":44,"value":407},{"type":39,"tag":128,"props":657,"children":658},{},[659,667,727],{"type":39,"tag":150,"props":660,"children":661},{},[662],{"type":39,"tag":65,"props":663,"children":665},{"className":664},[],[666],{"type":44,"value":267},{"type":39,"tag":150,"props":668,"children":669},{},[670,676,677,683,684,690,691,697,698,704,705,711,712,718,719,725],{"type":39,"tag":65,"props":671,"children":673},{"className":672},[],[674],{"type":44,"value":675},"codecategories",{"type":44,"value":201},{"type":39,"tag":65,"props":678,"children":680},{"className":679},[],[681],{"type":44,"value":682},"uicomponent",{"type":44,"value":201},{"type":39,"tag":65,"props":685,"children":687},{"className":686},[],[688],{"type":44,"value":689},"gamedev",{"type":44,"value":201},{"type":39,"tag":65,"props":692,"children":694},{"className":693},[],[695],{"type":44,"value":696},"3d",{"type":44,"value":201},{"type":39,"tag":65,"props":699,"children":701},{"className":700},[],[702],{"type":44,"value":703},"dataviz",{"type":44,"value":201},{"type":39,"tag":65,"props":706,"children":708},{"className":707},[],[709],{"type":44,"value":710},"image",{"type":44,"value":201},{"type":39,"tag":65,"props":713,"children":715},{"className":714},[],[716],{"type":44,"value":717},"video",{"type":44,"value":201},{"type":39,"tag":65,"props":720,"children":722},{"className":721},[],[723],{"type":44,"value":724},"svg",{"type":44,"value":726},", etc.",{"type":39,"tag":150,"props":728,"children":729},{},[730],{"type":44,"value":731},"Design Arena only.",{"type":39,"tag":128,"props":733,"children":734},{},[735,744,749],{"type":39,"tag":150,"props":736,"children":737},{},[738],{"type":39,"tag":65,"props":739,"children":741},{"className":740},[],[742],{"type":44,"value":743},"max_results",{"type":39,"tag":150,"props":745,"children":746},{},[747],{"type":44,"value":748},"positive integer",{"type":39,"tag":150,"props":750,"children":751},{},[752],{"type":44,"value":753},"Maximum number of rows returned by the API.",{"type":39,"tag":47,"props":755,"children":756},{},[757,759,765,766,772,774,780],{"type":44,"value":758},"Always preserve ",{"type":39,"tag":65,"props":760,"children":762},{"className":761},[],[763],{"type":44,"value":764},"meta.citation",{"type":44,"value":201},{"type":39,"tag":65,"props":767,"children":769},{"className":768},[],[770],{"type":44,"value":771},"meta.source_url",{"type":44,"value":773},", and ",{"type":39,"tag":65,"props":775,"children":777},{"className":776},[],[778],{"type":44,"value":779},"meta.as_of",{"type":44,"value":781},"; include attribution when republishing benchmark data.",{"type":39,"tag":47,"props":783,"children":784},{},[785,787,792,794,799],{"type":44,"value":786},"When results include both sources, do not present them as a single absolute leaderboard: Artificial Analysis indices and Design Arena ELO use different scales. Compare within each source, or rerun with ",{"type":39,"tag":65,"props":788,"children":790},{"className":789},[],[791],{"type":44,"value":233},{"type":44,"value":793}," or ",{"type":39,"tag":65,"props":795,"children":797},{"className":796},[],[798],{"type":44,"value":251},{"type":44,"value":800}," for a source-specific ranking.",{"type":39,"tag":53,"props":802,"children":804},{"id":803},"interpreting-results",[805],{"type":44,"value":806},"Interpreting Results",{"type":39,"tag":808,"props":809,"children":810},"ul",{},[811,836,895,913,930],{"type":39,"tag":380,"props":812,"children":813},{},[814,816,821,822,828,829,834],{"type":44,"value":815},"Artificial Analysis rows include ",{"type":39,"tag":65,"props":817,"children":819},{"className":818},[],[820],{"type":44,"value":340},{"type":44,"value":201},{"type":39,"tag":65,"props":823,"children":825},{"className":824},[],[826],{"type":44,"value":827},"coding_index",{"type":44,"value":773},{"type":39,"tag":65,"props":830,"children":832},{"className":831},[],[833],{"type":44,"value":348},{"type":44,"value":835},"; higher is better.",{"type":39,"tag":380,"props":837,"children":838},{},[839,841,847,848,854,855,861,862,867,868,873,874,880,882,887,888,893],{"type":44,"value":840},"Design Arena rows include ",{"type":39,"tag":65,"props":842,"children":844},{"className":843},[],[845],{"type":44,"value":846},"elo",{"type":44,"value":201},{"type":39,"tag":65,"props":849,"children":851},{"className":850},[],[852],{"type":44,"value":853},"win_rate",{"type":44,"value":201},{"type":39,"tag":65,"props":856,"children":858},{"className":857},[],[859],{"type":44,"value":860},"avg_generation_time_ms",{"type":44,"value":201},{"type":39,"tag":65,"props":863,"children":865},{"className":864},[],[866],{"type":44,"value":259},{"type":44,"value":201},{"type":39,"tag":65,"props":869,"children":871},{"className":870},[],[872],{"type":44,"value":267},{"type":44,"value":773},{"type":39,"tag":65,"props":875,"children":877},{"className":876},[],[878],{"type":44,"value":879},"tournament_stats",{"type":44,"value":881},"; higher ",{"type":39,"tag":65,"props":883,"children":885},{"className":884},[],[886],{"type":44,"value":846},{"type":44,"value":290},{"type":39,"tag":65,"props":889,"children":891},{"className":890},[],[892],{"type":44,"value":853},{"type":44,"value":894}," is better, lower generation time is faster.",{"type":39,"tag":380,"props":896,"children":897},{},[898,904,905,911],{"type":39,"tag":65,"props":899,"children":901},{"className":900},[],[902],{"type":44,"value":903},"pricing.prompt",{"type":44,"value":261},{"type":39,"tag":65,"props":906,"children":908},{"className":907},[],[909],{"type":44,"value":910},"pricing.completion",{"type":44,"value":912}," are USD per token as decimal strings. Multiply by 1,000,000 for per-million-token costs.",{"type":39,"tag":380,"props":914,"children":915},{},[916,921,923,928],{"type":39,"tag":65,"props":917,"children":919},{"className":918},[],[920],{"type":44,"value":367},{"type":44,"value":922}," identifies the benchmarked model entry. Verify it against ",{"type":39,"tag":65,"props":924,"children":926},{"className":925},[],[927],{"type":44,"value":390},{"type":44,"value":929}," before using it as a chat\u002Fcompletions model ID.",{"type":39,"tag":380,"props":931,"children":932},{},[933,939,941,947],{"type":39,"tag":65,"props":934,"children":936},{"className":935},[],[937],{"type":44,"value":938},"meta.model_count",{"type":44,"value":940}," counts unique models in the response, which can differ from ",{"type":39,"tag":65,"props":942,"children":944},{"className":943},[],[945],{"type":44,"value":946},"data.length",{"type":44,"value":948}," when multiple Design Arena categories are returned.",{"type":39,"tag":53,"props":950,"children":952},{"id":951},"direct-api-call",[953],{"type":44,"value":954},"Direct API Call",{"type":39,"tag":74,"props":956,"children":958},{"className":76,"code":957,"language":78,"meta":79,"style":79},"curl 'https:\u002F\u002Fopenrouter.ai\u002Fapi\u002Fv1\u002Fbenchmarks?source=artificial-analysis&task_type=coding&max_results=10' \\\n  -H \"Authorization: Bearer $OPENROUTER_API_KEY\"\n",[959],{"type":39,"tag":65,"props":960,"children":961},{"__ignoreMap":79},[962,992],{"type":39,"tag":85,"props":963,"children":964},{"class":87,"line":88},[965,971,976,982,987],{"type":39,"tag":85,"props":966,"children":968},{"style":967},"--shiki-light:#E2931D;--shiki-default:#FFCB6B;--shiki-dark:#FFCB6B",[969],{"type":44,"value":970},"curl",{"type":39,"tag":85,"props":972,"children":973},{"style":104},[974],{"type":44,"value":975}," '",{"type":39,"tag":85,"props":977,"children":979},{"style":978},"--shiki-light:#91B859;--shiki-default:#C3E88D;--shiki-dark:#C3E88D",[980],{"type":44,"value":981},"https:\u002F\u002Fopenrouter.ai\u002Fapi\u002Fv1\u002Fbenchmarks?source=artificial-analysis&task_type=coding&max_results=10",{"type":39,"tag":85,"props":983,"children":984},{"style":104},[985],{"type":44,"value":986},"'",{"type":39,"tag":85,"props":988,"children":989},{"style":98},[990],{"type":44,"value":991}," \\\n",{"type":39,"tag":85,"props":993,"children":995},{"class":87,"line":994},2,[996,1001,1006,1011,1016],{"type":39,"tag":85,"props":997,"children":998},{"style":978},[999],{"type":44,"value":1000},"  -H",{"type":39,"tag":85,"props":1002,"children":1003},{"style":104},[1004],{"type":44,"value":1005}," \"",{"type":39,"tag":85,"props":1007,"children":1008},{"style":978},[1009],{"type":44,"value":1010},"Authorization: Bearer ",{"type":39,"tag":85,"props":1012,"children":1013},{"style":98},[1014],{"type":44,"value":1015},"$OPENROUTER_API_KEY",{"type":39,"tag":85,"props":1017,"children":1018},{"style":104},[1019],{"type":44,"value":1020},"\"\n",{"type":39,"tag":47,"props":1022,"children":1023},{},[1024,1025,1030],{"type":44,"value":311},{"type":39,"tag":65,"props":1026,"children":1028},{"className":1027},[],[1029],{"type":44,"value":317},{"type":44,"value":1031}," when implementing against the raw API or handling source-specific response shapes.",{"type":39,"tag":1033,"props":1034,"children":1035},"style",{},[1036],{"type":44,"value":1037},"html .light .shiki span {color: var(--shiki-light);background: var(--shiki-light-bg);font-style: var(--shiki-light-font-style);font-weight: var(--shiki-light-font-weight);text-decoration: var(--shiki-light-text-decoration);}html.light .shiki span {color: var(--shiki-light);background: var(--shiki-light-bg);font-style: var(--shiki-light-font-style);font-weight: var(--shiki-light-font-weight);text-decoration: var(--shiki-light-text-decoration);}html .default .shiki span {color: var(--shiki-default);background: var(--shiki-default-bg);font-style: var(--shiki-default-font-style);font-weight: var(--shiki-default-font-weight);text-decoration: var(--shiki-default-text-decoration);}html .shiki span {color: var(--shiki-default);background: var(--shiki-default-bg);font-style: var(--shiki-default-font-style);font-weight: var(--shiki-default-font-weight);text-decoration: var(--shiki-default-text-decoration);}html .dark .shiki span {color: var(--shiki-dark);background: var(--shiki-dark-bg);font-style: var(--shiki-dark-font-style);font-weight: var(--shiki-dark-font-weight);text-decoration: var(--shiki-dark-text-decoration);}html.dark .shiki span {color: var(--shiki-dark);background: var(--shiki-dark-bg);font-style: var(--shiki-dark-font-style);font-weight: var(--shiki-dark-font-weight);text-decoration: var(--shiki-dark-text-decoration);}",{"items":1039,"total":1124},[1040,1056,1066,1080,1094,1103,1112],{"slug":1041,"name":1041,"fn":1042,"description":1043,"org":1044,"tags":1045,"stars":23,"repoUrl":24,"updatedAt":1055},"create-agent-tui","scaffold terminal agents with OpenRouter","Scaffolds a complete agent TUI in TypeScript using @openrouter\u002Fagent — like create-react-app for terminal agents. Generates a customizable terminal interface with three input styles, four tool display modes, ASCII banners, streaming output, session persistence, and configurable tools. Use when building an agent, creating a TUI, scaffolding an agent project, or building a coding assistant.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":11},[1046,1048,1051,1052],{"name":1047,"slug":644,"type":16},"Agents",{"name":1049,"slug":1050,"type":16},"CLI","cli",{"name":9,"slug":8,"type":16},{"name":1053,"slug":1054,"type":16},"TypeScript","typescript","2026-07-14T05:38:18.849741",{"slug":1057,"name":1057,"fn":1058,"description":1059,"org":1060,"tags":1061,"stars":23,"repoUrl":24,"updatedAt":1065},"create-headless-agent","scaffold headless agents with OpenRouter","Scaffolds a headless agent in TypeScript using @openrouter\u002Fagent and Bun — for CLI tools, API servers, queue workers, and pipelines. No terminal UI. Use when building a headless agent, programmatic agent, CLI tool that uses AI, batch agent, pipeline agent, API agent, agent without a UI, or agent service.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":11},[1062,1063,1064],{"name":1047,"slug":644,"type":16},{"name":1049,"slug":1050,"type":16},{"name":1053,"slug":1054,"type":16},"2026-07-14T05:38:30.178029",{"slug":1067,"name":1067,"fn":1068,"description":1069,"org":1070,"tags":1071,"stars":23,"repoUrl":24,"updatedAt":1079},"open-responses","implement Open Responses-compliant APIs","This skill should be used when implementing, consuming, or debugging an Open Responses-compliant API — the open standard for multi-provider LLM interoperability. Covers protocol, items, state machines, streaming events, tools, the agentic loop pattern, and extensions. Triggers on: Open Responses, open-responses, \u002Fv1\u002Fresponses endpoint, multi-provider LLM API, Open Responses compliance.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":11},[1072,1075,1078],{"name":1073,"slug":1074,"type":16},"API Development","api-development",{"name":1076,"slug":1077,"type":16},"Interoperability","interoperability",{"name":18,"slug":19,"type":16},"2026-07-14T05:38:13.153467",{"slug":1081,"name":1081,"fn":1082,"description":1083,"org":1084,"tags":1085,"stars":23,"repoUrl":24,"updatedAt":1093},"openrouter-agent-migration","migrate OpenRouter SDK to agent patterns","Migration guide from @openrouter\u002Fsdk to @openrouter\u002Fagent for callModel, tool(), stop conditions, and agent features. This skill should be used when code imports callModel, tool(), or stop conditions from @openrouter\u002Fsdk and needs to migrate to @openrouter\u002Fagent.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":11},[1086,1089,1092],{"name":1087,"slug":1088,"type":16},"Migration","migration",{"name":1090,"slug":1091,"type":16},"SDK","sdk",{"name":1053,"slug":1054,"type":16},"2026-07-14T05:38:28.914981",{"slug":1095,"name":1095,"fn":1096,"description":1097,"org":1098,"tags":1099,"stars":23,"repoUrl":24,"updatedAt":1102},"openrouter-analytics","analyze OpenRouter usage and spend","Answer natural-language questions about a user's OpenRouter usage data — spend, request volume, model breakdown, latency, token usage, and cost optimization. Use when the user asks about their API usage, billing, costs, top models, traffic patterns, or wants to optimize their OpenRouter spend.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":11},[1100,1101],{"name":21,"slug":22,"type":16},{"name":9,"slug":8,"type":16},"2026-07-30T05:30:15.098344",{"slug":1104,"name":1104,"fn":1105,"description":1106,"org":1107,"tags":1108,"stars":23,"repoUrl":24,"updatedAt":1111},"openrouter-analytics-query","execute analytics queries against OpenRouter","Construct and execute analytics queries against the OpenRouter API — full parameter reference for metrics, dimensions, filters, time ranges, ordering, and pagination. Use when building or debugging an analytics query, understanding the request\u002Fresponse shape, or handling query errors.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":11},[1109,1110],{"name":21,"slug":22,"type":16},{"name":1073,"slug":1074,"type":16},"2026-07-14T05:38:26.402047",{"slug":1113,"name":1113,"fn":1114,"description":1115,"org":1116,"tags":1117,"stars":23,"repoUrl":24,"updatedAt":1123},"openrouter-analytics-schema","query OpenRouter analytics schema","Discover the OpenRouter analytics schema — available metrics, dimensions, filter operators, and granularities. Use when you need to know what analytics data is queryable, what dimensions you can break down by, or how to map a user's question to the right metric\u002Fdimension combination.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":11},[1118,1119,1120],{"name":21,"slug":22,"type":16},{"name":18,"slug":19,"type":16},{"name":1121,"slug":1122,"type":16},"Reporting","reporting","2026-07-14T05:38:32.721796",17,{"items":1126,"total":1124},[1127,1134,1140,1146,1152,1157,1162,1168,1174,1183,1194,1202],{"slug":1041,"name":1041,"fn":1042,"description":1043,"org":1128,"tags":1129,"stars":23,"repoUrl":24,"updatedAt":1055},{"slug":8,"name":9,"logoUrl":10,"githubOrg":11},[1130,1131,1132,1133],{"name":1047,"slug":644,"type":16},{"name":1049,"slug":1050,"type":16},{"name":9,"slug":8,"type":16},{"name":1053,"slug":1054,"type":16},{"slug":1057,"name":1057,"fn":1058,"description":1059,"org":1135,"tags":1136,"stars":23,"repoUrl":24,"updatedAt":1065},{"slug":8,"name":9,"logoUrl":10,"githubOrg":11},[1137,1138,1139],{"name":1047,"slug":644,"type":16},{"name":1049,"slug":1050,"type":16},{"name":1053,"slug":1054,"type":16},{"slug":1067,"name":1067,"fn":1068,"description":1069,"org":1141,"tags":1142,"stars":23,"repoUrl":24,"updatedAt":1079},{"slug":8,"name":9,"logoUrl":10,"githubOrg":11},[1143,1144,1145],{"name":1073,"slug":1074,"type":16},{"name":1076,"slug":1077,"type":16},{"name":18,"slug":19,"type":16},{"slug":1081,"name":1081,"fn":1082,"description":1083,"org":1147,"tags":1148,"stars":23,"repoUrl":24,"updatedAt":1093},{"slug":8,"name":9,"logoUrl":10,"githubOrg":11},[1149,1150,1151],{"name":1087,"slug":1088,"type":16},{"name":1090,"slug":1091,"type":16},{"name":1053,"slug":1054,"type":16},{"slug":1095,"name":1095,"fn":1096,"description":1097,"org":1153,"tags":1154,"stars":23,"repoUrl":24,"updatedAt":1102},{"slug":8,"name":9,"logoUrl":10,"githubOrg":11},[1155,1156],{"name":21,"slug":22,"type":16},{"name":9,"slug":8,"type":16},{"slug":1104,"name":1104,"fn":1105,"description":1106,"org":1158,"tags":1159,"stars":23,"repoUrl":24,"updatedAt":1111},{"slug":8,"name":9,"logoUrl":10,"githubOrg":11},[1160,1161],{"name":21,"slug":22,"type":16},{"name":1073,"slug":1074,"type":16},{"slug":1113,"name":1113,"fn":1114,"description":1115,"org":1163,"tags":1164,"stars":23,"repoUrl":24,"updatedAt":1123},{"slug":8,"name":9,"logoUrl":10,"githubOrg":11},[1165,1166,1167],{"name":21,"slug":22,"type":16},{"name":18,"slug":19,"type":16},{"name":1121,"slug":1122,"type":16},{"slug":4,"name":4,"fn":5,"description":6,"org":1169,"tags":1170,"stars":23,"repoUrl":24,"updatedAt":25},{"slug":8,"name":9,"logoUrl":10,"githubOrg":11},[1171,1172,1173],{"name":21,"slug":22,"type":16},{"name":14,"slug":15,"type":16},{"name":18,"slug":19,"type":16},{"slug":1175,"name":1175,"fn":1176,"description":1177,"org":1178,"tags":1179,"stars":23,"repoUrl":24,"updatedAt":1182},"openrouter-generations","retrieve metadata for OpenRouter generations","Retrieve detailed metadata and stored content for individual OpenRouter generations. Use when the user wants to inspect a specific request — its cost, latency, token usage, provider routing, or the actual prompt\u002Fcompletion text — or is debugging a failed or unexpected generation.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":11},[1180,1181],{"name":18,"slug":19,"type":16},{"name":9,"slug":8,"type":16},"2026-07-14T05:38:25.132801",{"slug":1184,"name":1184,"fn":1185,"description":1186,"org":1187,"tags":1188,"stars":23,"repoUrl":24,"updatedAt":1193},"openrouter-images","generate images with OpenRouter","Generate images from text prompts and edit existing images using OpenRouter's dedicated Image API. Use when the user asks to create, generate, or make an image, picture, or illustration from a description, or wants to edit, modify, transform, or alter an existing image with a text prompt.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":11},[1189,1192],{"name":1190,"slug":1191,"type":16},"Image Generation","image-generation",{"name":9,"slug":8,"type":16},"2026-07-14T05:38:21.393411",{"slug":327,"name":327,"fn":1195,"description":1196,"org":1197,"tags":1198,"stars":23,"repoUrl":24,"updatedAt":1201},"query OpenRouter model metadata and pricing","Query OpenRouter for available AI models, pricing, capabilities, throughput, and provider performance. Use when the user asks about available OpenRouter models, model pricing, model context lengths, model capabilities, provider latency or uptime, throughput limits, supported parameters, wants to search\u002Ffilter\u002Fcompare models, or find the fastest provider for a model.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":11},[1199,1200],{"name":18,"slug":19,"type":16},{"name":9,"slug":8,"type":16},"2026-07-14T05:38:22.650307",{"slug":1203,"name":1203,"fn":1204,"description":1205,"org":1206,"tags":1207,"stars":23,"repoUrl":24,"updatedAt":1215},"openrouter-oauth","implement OpenRouter OAuth authentication","Implement \"Sign In with OpenRouter\" using OAuth PKCE — framework-agnostic, no SDK or client registration required. Use when the user wants to add OpenRouter login, authentication, sign-in buttons, OAuth, or AI model inference API keys for browser-based apps. No client registration, no backend, no secrets required.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":11},[1208,1211,1214],{"name":1209,"slug":1210,"type":16},"Auth","auth",{"name":1212,"slug":1213,"type":16},"OAuth","oauth",{"name":9,"slug":8,"type":16},"2026-07-14T05:38:20.116308"]