{
  "listings": [
    {
      "slug": "azure-openai",
      "name": "Azure OpenAI in Foundry Models",
      "company": "Microsoft Corporation",
      "grade": "A",
      "fit": "yes",
      "why": "response_format json_schema with strict true and strict function tools, refusal and finish_reason fields on every response; schema limits are the OpenAI subset.",
      "source": "https://learn.microsoft.com/en-us/azure/foundry/openai/how-to/structured-outputs",
      "confidence": "medium",
      "url": "https://www.anchorterminal.com/tools/azure-openai"
    },
    {
      "slug": "cerebras-inference",
      "name": "Cerebras Inference",
      "company": "Cerebras Systems Inc.",
      "grade": "B",
      "fit": "yes",
      "why": "response_format json_schema with strict true uses constrained decoding on the listed models, every object needing additionalProperties false; strict false is a hint only.",
      "source": "https://inference-docs.cerebras.ai/capabilities/structured-outputs",
      "confidence": "high",
      "url": "https://www.anchorterminal.com/tools/cerebras-inference"
    },
    {
      "slug": "cohere-chat",
      "name": "Cohere Chat API (Command models)",
      "company": "Cohere Inc.",
      "grade": "B",
      "fit": "yes",
      "why": "response_format json_object or json_schema forces the schema and strict_tools does the same for tool arguments; not available in RAG mode and json_object needs an explicit JSON instruction or the model can loop.",
      "source": "https://docs.cohere.com/docs/structured-outputs",
      "confidence": "medium",
      "url": "https://www.anchorterminal.com/tools/cohere-chat"
    },
    {
      "slug": "deepinfra",
      "name": "DeepInfra",
      "company": "Deep Infra Inc.",
      "grade": "B",
      "fit": "yes",
      "why": "response_format json_schema with strict constrains output to the schema on the models tagged for it; no refusal field is described.",
      "source": "https://docs.deepinfra.com/chat/structured-outputs",
      "confidence": "medium",
      "url": "https://www.anchorterminal.com/tools/deepinfra"
    },
    {
      "slug": "fireworks-inference",
      "name": "Fireworks AI Inference",
      "company": "Fireworks.ai, Inc.",
      "grade": "BB",
      "fit": "yes",
      "why": "response_format json_schema or a grammar constrains output, finish_reason length flags truncation, and the supported and unsupported schema keywords are listed; pattern is best effort.",
      "source": "https://docs.fireworks.ai/structured-responses/structured-response-formatting",
      "confidence": "medium",
      "url": "https://www.anchorterminal.com/tools/fireworks-inference"
    },
    {
      "slug": "meta-model-api",
      "name": "Meta Model API",
      "company": "Meta Platforms, Inc.",
      "grade": "D",
      "fit": "yes",
      "why": "response_format json_schema constrains decoding so output always matches, recursive schemas return 400, and a 120,000-character schema limit is stated; strict false behaves as true for response_format.",
      "source": "https://dev.meta.ai/docs/structured-output",
      "confidence": "medium",
      "url": "https://www.anchorterminal.com/tools/meta-model-api"
    },
    {
      "slug": "amazon-bedrock",
      "name": "Amazon Bedrock",
      "company": "Amazon Web Services, Inc.",
      "grade": "A",
      "fit": "caveat",
      "why": "Strict tool use and output_config.format validate against a JSON Schema 2020-12 subset server-side, but a new schema compiles for up to 90 seconds, every object needs additionalProperties false, and schemas are capped at 12 optional and 16 union-typed properties.",
      "source": "https://docs.aws.amazon.com/bedrock/latest/userguide/structured-output.html",
      "confidence": "medium",
      "url": "https://www.anchorterminal.com/tools/amazon-bedrock"
    },
    {
      "slug": "anthropic-api",
      "name": "Claude API",
      "company": "Anthropic, PBC",
      "grade": "A",
      "fit": "caveat",
      "why": "output_config.format json_schema and strict tools use a server grammar, with stop_reason refusal and max_tokens as signals. The docs admit an enum value differing only in case can pass with no error, so caller validation is still needed.",
      "source": "https://platform.claude.com/docs/en/build-with-claude/structured-outputs",
      "confidence": "high",
      "url": "https://www.anchorterminal.com/tools/anthropic-api"
    },
    {
      "slug": "blockrun-ai",
      "name": "BlockRun.AI",
      "company": "BlockRun, Inc.",
      "grade": "A",
      "fit": "caveat",
      "why": "response_format accepts json_object and json_schema and is forwarded to the upstream lab; the OpenAPI document has no strict field and no refusal path, so enforcement is whatever the upstream does.",
      "source": "https://blockrun.ai/docs/api-reference/chat-completions",
      "confidence": "medium",
      "url": "https://www.anchorterminal.com/tools/blockrun-ai"
    },
    {
      "slug": "cloudflare-ai-gateway",
      "name": "Cloudflare AI Gateway",
      "company": "Cloudflare, Inc.",
      "grade": "B",
      "fit": "caveat",
      "why": "No schema enforcement of its own; response_format passes through to the provider, so validity is the upstream's.",
      "source": "https://developers.cloudflare.com/ai-gateway/usage/chat-completion/",
      "confidence": "medium",
      "url": "https://www.anchorterminal.com/tools/cloudflare-ai-gateway"
    },
    {
      "slug": "cloudflare-workers-ai",
      "name": "Cloudflare Workers AI",
      "company": "Cloudflare, Inc.",
      "grade": "BB",
      "fit": "caveat",
      "why": "JSON mode takes response_format json_schema on four listed models, one of which (llama-3.1-8b-instruct) is on the retirement list; no strict flag or refusal path is documented.",
      "source": "https://developers.cloudflare.com/workers-ai/features/json-mode/",
      "confidence": "medium",
      "url": "https://www.anchorterminal.com/tools/cloudflare-workers-ai"
    },
    {
      "slug": "deepseek-api",
      "name": "DeepSeek API",
      "company": "Hangzhou DeepSeek Artificial Intelligence Co., Ltd.",
      "grade": "D",
      "fit": "caveat",
      "why": "response_format has json_object only, the caller must also ask for JSON in the prompt, and non-JSON output is replaced with spaces so the content comes back blank; schema enforcement exists only for tool arguments in beta strict mode.",
      "source": "https://api-docs.deepseek.com/guides/json_mode",
      "confidence": "high",
      "url": "https://www.anchorterminal.com/tools/deepseek-api"
    },
    {
      "slug": "eden-ai",
      "name": "Eden AI",
      "company": "Eden AI SAS",
      "grade": "C",
      "fit": "caveat",
      "why": "response_format is forwarded and a json_heal hook can repair output against a json_schema after the call, but no strict flag or refusal field is documented.",
      "source": "https://www.edenai.co/docs/api-reference/chat/chat-completions",
      "confidence": "medium",
      "url": "https://www.anchorterminal.com/tools/eden-ai"
    },
    {
      "slug": "elorian-glimpse",
      "name": "Glimpse",
      "company": "Elorian AI, Inc.",
      "grade": "F",
      "fit": "caveat",
      "why": "response_format and structured_outputs are in the endpoint's supported parameters and OpenRouter forwards strict json_schema, but enforcement is Together's and no refusal path is documented.",
      "source": "https://openrouter.ai/api/v1/models/elorian/glimpse/endpoints",
      "confidence": "low",
      "url": "https://www.anchorterminal.com/tools/elorian-glimpse"
    },
    {
      "slug": "gemini-api",
      "name": "Gemini Developer API",
      "company": "Google LLC",
      "grade": "B",
      "fit": "caveat",
      "why": "response_format json_schema constrains decoding, but several keywords (minimum, maxContains and others) are ignored and left to the client, and acc-06 saw one model wrap JSON in Markdown fences while Flash behaved.",
      "source": "https://aistudio.google.com/docs/structured-output",
      "confidence": "high",
      "url": "https://www.anchorterminal.com/tools/gemini-api"
    },
    {
      "slug": "groq",
      "name": "GroqCloud",
      "company": "Groq LLC",
      "grade": "BB",
      "fit": "caveat",
      "why": "strict true uses constrained decoding and guarantees the schema, but only on select models; best-effort mode can return malformed JSON or 400 with failed_generation.",
      "source": "https://console.groq.com/docs/structured-outputs",
      "confidence": "high",
      "url": "https://www.anchorterminal.com/tools/groq"
    },
    {
      "slug": "hugging-face-inference-providers",
      "name": "Hugging Face Inference Providers",
      "company": "Hugging Face, Inc.",
      "grade": "C",
      "fit": "caveat",
      "why": "response_format json_schema is forwarded and the guide promises schema compliance, but supports_structured_output is true on only 102 of 247 provider endpoints and no refusal path is documented.",
      "source": "https://huggingface.co/docs/inference-providers/guides/structured-output",
      "confidence": "medium",
      "url": "https://www.anchorterminal.com/tools/hugging-face-inference-providers"
    },
    {
      "slug": "litellm",
      "name": "LiteLLM",
      "company": "Berrie AI Incorporated",
      "grade": "B",
      "fit": "caveat",
      "why": "response_format json_schema is forwarded and get_supported_openai_params says whether a model takes it; enforcement is the provider's.",
      "source": "https://docs.litellm.ai/docs/completion/json_mode",
      "confidence": "high",
      "url": "https://www.anchorterminal.com/tools/litellm"
    },
    {
      "slug": "mistral-api",
      "name": "Mistral AI API",
      "company": "Mistral AI (RCS Paris 952 418 325)",
      "grade": "B",
      "fit": "caveat",
      "why": "response_format json_schema with strict true on all models except codestral-mamba; no refusal field is documented and the known-limitations page says JSON mode alone doesn't follow a schema.",
      "source": "https://docs.mistral.ai/inference/conversations/structured-output/custom.md",
      "confidence": "medium",
      "url": "https://www.anchorterminal.com/tools/mistral-api"
    },
    {
      "slug": "muse-spark-open",
      "name": "Muse Spark Open",
      "company": "Meta Platforms, Inc.",
      "grade": "C",
      "fit": "caveat",
      "why": "Meta's examples use response_format json_object on a hosted provider; schema enforcement depends on the serving stack and no strict flag or refusal path is documented.",
      "source": "https://dev.meta.ai/docs/muse-spark-open/deploy.md",
      "confidence": "medium",
      "url": "https://www.anchorterminal.com/tools/muse-spark-open"
    },
    {
      "slug": "nemotron-3-5-super-vl",
      "name": "Nemotron 3.5 Super VL",
      "company": "NVIDIA Corporation",
      "grade": "C",
      "fit": "caveat",
      "why": "response_format accepts json_object and json_schema; no strict flag, refusal field or validation statement in the reference.",
      "source": "https://docs.api.nvidia.com/nim/reference/nvidia-nemotron-3_5-super-vl-120b-a12b-infer.md",
      "confidence": "medium",
      "url": "https://www.anchorterminal.com/tools/nemotron-3-5-super-vl"
    },
    {
      "slug": "novita-ai",
      "name": "Novita AI",
      "company": "Novita AI",
      "grade": "C",
      "fit": "caveat",
      "why": "json_schema response_format 'ensures the model will match your supplied JSON schema' on listed models; no strict flag or refusal field, and the listing's open question notes a FAQ that calls json_schema unsupported.",
      "source": "https://novita.ai/docs/api-reference/model-apis-llm-create-chat-completion.md",
      "confidence": "medium",
      "url": "https://www.anchorterminal.com/tools/novita-ai"
    },
    {
      "slug": "openai-api",
      "name": "OpenAI API",
      "company": "OpenAI OpCo, LLC",
      "grade": "AA",
      "fit": "caveat",
      "why": "strict json_schema with a refusal content type and an incomplete status reason of max_output_tokens; acc-05 reports a schema that validated for weeks failing on 18 June 2026 with no changelog.",
      "source": "https://developers.openai.com/api/docs/guides/structured-outputs.md",
      "confidence": "high",
      "url": "https://www.anchorterminal.com/tools/openai-api"
    },
    {
      "slug": "openrouter",
      "name": "OpenRouter",
      "company": "OpenRouter, Inc.",
      "grade": "B",
      "fit": "caveat",
      "why": "json_schema with strict true, but enforcement varies by provider; without require_parameters a request may land on a host that ignores it or fail for lack of support.",
      "source": "https://openrouter.ai/docs/guides/features/structured-outputs.md",
      "confidence": "high",
      "url": "https://www.anchorterminal.com/tools/openrouter"
    },
    {
      "slug": "prism-inference",
      "name": "Prism Inference",
      "company": "Prism Technologies Inc",
      "grade": "C",
      "fit": "caveat",
      "why": "JSON mode and strict JSON Schema on Chat Completions, with strict in the OpenAPI; no refusal field or validation statement, and no first-hand account.",
      "source": "https://docs.prisminference.com/models.md",
      "confidence": "medium",
      "url": "https://www.anchorterminal.com/tools/prism-inference"
    },
    {
      "slug": "ramp-router",
      "name": "Router.com (Ramp Router)",
      "company": "Ramp Business Corporation",
      "grade": "C",
      "fit": "caveat",
      "why": "response_format json_object is translated natively; a strict json_schema path isn't shown in the pages read, and model refusals end the request without fallback.",
      "source": "https://docs.router.com/api/responses-and-messages.md",
      "confidence": "medium",
      "url": "https://www.anchorterminal.com/tools/ramp-router"
    },
    {
      "slug": "reflection-beam",
      "name": "Beam 1.0",
      "company": "Reflection AI",
      "grade": "C",
      "fit": "caveat",
      "why": "json_schema response_format with a strict mode, a refusal field and finish_reason length documented; there is no changelog, so schema stability over time can't be checked.",
      "source": "https://developers.reflection.ai/structured-outputs.md",
      "confidence": "medium",
      "url": "https://www.anchorterminal.com/tools/reflection-beam"
    },
    {
      "slug": "routerplus",
      "name": "RouterPlus",
      "company": "EVO HQ Inc",
      "grade": "D",
      "fit": "caveat",
      "why": "response_format works for GPT ids and open models, not Claude; on /v1/messages output_config.format maps to a strict json_schema for open models; no refusal path documented.",
      "source": "https://app.routerplus.com/llms-full.txt",
      "confidence": "medium",
      "url": "https://www.anchorterminal.com/tools/routerplus"
    },
    {
      "slug": "sambanova",
      "name": "SambaCloud",
      "company": "SambaNova Systems, Inc.",
      "grade": "B",
      "fit": "caveat",
      "why": "json_schema response_format on supported models, but strict is accepted for true and false and only enforced 'when strict enforcement is available', so validation stays with the caller.",
      "source": "https://docs.sambanova.ai/docs/en/features/function-calling.md",
      "confidence": "medium",
      "url": "https://www.anchorterminal.com/tools/sambanova"
    },
    {
      "slug": "siliconflow",
      "name": "SiliconFlow",
      "company": "SiliconFlow Labs Pte. Ltd.",
      "grade": "E",
      "fit": "caveat",
      "why": "JSON Mode (json_object) only, on most models except DeepSeek R1 and V3; the guide says the app must handle incomplete JSON, and there is no json_schema or strict.",
      "source": "https://docs.siliconflow.com/en/userguide/guides/json-mode.md",
      "confidence": "medium",
      "url": "https://www.anchorterminal.com/tools/siliconflow"
    },
    {
      "slug": "together-inference",
      "name": "Together AI Inference",
      "company": "Together Computer, Inc.",
      "grade": "B",
      "fit": "caveat",
      "why": "json_schema response_format on supported models, but the guide pairs it with prompt instructions and documents no strict flag or refusal field.",
      "source": "https://docs.together.ai/docs/inference/chat/structured-outputs.md",
      "confidence": "medium",
      "url": "https://www.anchorterminal.com/tools/together-inference"
    },
    {
      "slug": "xai-api",
      "name": "xAI API (Grok)",
      "company": "SpaceXSI (xAI)",
      "grade": "BB",
      "fit": "caveat",
      "why": "json_schema response_format with strict true; structured outputs with tools only on supported Grok 4 models, and no refusal field in the pages read.",
      "source": "https://docs.x.ai/llms-full.txt",
      "confidence": "medium",
      "url": "https://www.anchorterminal.com/tools/xai-api"
    },
    {
      "slug": "antseed",
      "name": "Antseed",
      "company": "Antseed Foundation",
      "grade": "C",
      "fit": "unchecked",
      "why": "No response_format or schema enforcement appears in the API or routing pages read.",
      "confidence": "low",
      "url": "https://www.anchorterminal.com/tools/antseed"
    },
    {
      "slug": "jais-3",
      "name": "Jais 3",
      "company": "Mohamed bin Zayed University of Artificial Intelligence (Institute of Foundation Models), with Inception42 (a G42 company)",
      "grade": "C",
      "fit": "unchecked",
      "why": "Schema enforcement would come from vLLM or SGLang guided decoding; the cards are silent.",
      "confidence": "medium",
      "url": "https://www.anchorterminal.com/tools/jais-3"
    },
    {
      "slug": "stepfun-api",
      "name": "StepFun API",
      "company": "STEPFUN TECHNOLOGY (HK) LIMITED",
      "grade": "E",
      "fit": "unchecked",
      "why": "The host platform.stepfun.ai closed the connection on every attempt (curl exit 56, no HTTP status) from this session, so no page was read.",
      "confidence": "low",
      "url": "https://www.anchorterminal.com/tools/stepfun-api"
    },
    {
      "slug": "tencent-tokenhub",
      "name": "Tencent Cloud TokenHub",
      "company": "Tencent Cloud Computing (Beijing) Co., Ltd. (腾讯云计算（北京）有限责任公司)",
      "grade": "C",
      "fit": "unchecked",
      "why": "The host cloud.tencent.com closed the connection on every attempt (curl exit 56, no HTTP status) from this session, so no page was read.",
      "confidence": "low",
      "url": "https://www.anchorterminal.com/tools/tencent-tokenhub"
    },
    {
      "slug": "vercel-ai-gateway",
      "name": "Vercel AI Gateway",
      "company": "Vercel Inc.",
      "grade": "B",
      "fit": "unchecked",
      "why": "The structured output pages weren't opened; response_format passes through the OpenAI-compatible endpoint but strictness is the provider's.",
      "confidence": "medium",
      "url": "https://www.anchorterminal.com/tools/vercel-ai-gateway"
    },
    {
      "slug": "vertex-ai",
      "name": "Vertex AI (Gemini Enterprise Agent Platform)",
      "company": "Google LLC",
      "grade": "AA",
      "fit": "unchecked",
      "why": "The structured output page returned 404 at the paths tried, so response_schema wasn't verified here.",
      "confidence": "low",
      "url": "https://www.anchorterminal.com/tools/vertex-ai"
    },
    {
      "slug": "minimax-api",
      "name": "MiniMax API",
      "company": "Nanonoble Pte. Ltd.",
      "grade": "C",
      "fit": "no",
      "why": "Neither OpenAPI file has response_format or a JSON schema mode, and the OpenAI-format schema also lacks tool_choice, so schema enforcement is the caller's job.",
      "source": "https://platform.minimax.io/docs/api-reference/text/api/openapi-chat-openai.json",
      "confidence": "medium",
      "url": "https://www.anchorterminal.com/tools/minimax-api"
    },
    {
      "slug": "simplismart",
      "name": "Simplismart",
      "company": "Verute Technologies Private Limited",
      "grade": "F",
      "fit": "no",
      "why": "No response_format or JSON mode in the OpenAPI file.",
      "source": "https://docs.simplismart.ai/openapi-specs/playground/llama-8b.yaml",
      "confidence": "low",
      "url": "https://www.anchorterminal.com/tools/simplismart"
    },
    {
      "slug": "swe-2-flash",
      "name": "SWE-2 Flash",
      "company": "Cognition AI, Inc.",
      "grade": "F",
      "fit": "no",
      "why": "No API and no schema mode.",
      "source": "https://docs.devin.ai/desktop/models.md",
      "confidence": "medium",
      "url": "https://www.anchorterminal.com/tools/swe-2-flash"
    }
  ],
  "marks": "yes: the docs show it does the job as the success test says. caveat: it does the job with a limit an agent should know. no: it doesn't do this job. unchecked: we couldn't read it.",
  "meta": {
    "attribution": "Anchor Terminal (https://www.anchorterminal.com)",
    "attributionRequired": true,
    "credit": "CC BY 4.0, credit Anchor Terminal and link the page or API address you took it from (https://www.anchorterminal.com/terms/#data)",
    "docs": "https://www.anchorterminal.com/docs/",
    "generatedAt": "2026-11-16",
    "license": "CC-BY-4.0",
    "method": "https://www.anchorterminal.com/benchmark/",
    "methodology": "0.8",
    "openapi": "https://www.anchorterminal.com/openapi.json",
    "preview": false,
    "run": "2026-10-01",
    "runLabel": "October 2026 research run",
    "terms": "https://www.anchorterminal.com/terms/#data"
  },
  "useCase": {
    "accounts": 4,
    "after": [
      "vector-db.store",
      "document-extraction.table-from-text"
    ],
    "area": "models",
    "asks": [
      "extract these fields as JSON",
      "classify each row into one of these labels",
      "give me output my code can parse without cleaning"
    ],
    "before": [
      "web-scrapers.page-to-markdown",
      "document-extraction.pdf-to-text"
    ],
    "blockers": [
      "server-side schema validator changes with no changelog, rejecting schemas that passed",
      "model wraps JSON in Markdown fences on one model while siblings behave",
      "strict flag accepted but ignored, so validation is the caller's job",
      "stop sequences ignored, breaking the parser"
    ],
    "capabilities": [
      "inference.llm"
    ],
    "category": "inference",
    "categoryName": "Model APIs \u0026 inference",
    "checked": "2026-11-12",
    "companies": {
      "fit": 6,
      "fitWithLimit": 24,
      "tagged": 42
    },
    "competition": "crowded",
    "fits": [
      "azure-openai",
      "cerebras-inference",
      "cohere-chat",
      "deepinfra",
      "fireworks-inference",
      "meta-model-api"
    ],
    "fitsWithLimit": [
      "amazon-bedrock",
      "anthropic-api",
      "blockrun-ai",
      "cloudflare-ai-gateway",
      "cloudflare-workers-ai",
      "deepseek-api",
      "eden-ai",
      "elorian-glimpse",
      "gemini-api",
      "groq",
      "hugging-face-inference-providers",
      "litellm",
      "mistral-api",
      "muse-spark-open",
      "nemotron-3-5-super-vl",
      "novita-ai",
      "openai-api",
      "openrouter",
      "prism-inference",
      "ramp-router",
      "reflection-beam",
      "routerplus",
      "sambanova",
      "siliconflow",
      "together-inference",
      "xai-api"
    ],
    "id": "inference.structured-extraction",
    "listings": {
      "caveat": 26,
      "no": 3,
      "notRead": 6,
      "tagged": 51,
      "unchecked": 6,
      "yes": 6
    },
    "name": "Return JSON that matches a schema",
    "status": "draft",
    "success": "Output parses as JSON and validates against the schema, and refusals and truncation are signalled in a field the caller can check. Measured: output parses and validates on every call in a 100-call run, and a schema that validated last week still validates this week.",
    "url": "https://www.anchorterminal.com/api/v1/use-cases/inference.structured-extraction.json",
    "weight": 10,
    "weightBasis": "4 of 28 accounts; two are forum threads with a dozen users each reporting valid schemas that stopped validating"
  }
}
