{"generatedBy":"scripts/providerConformance.ts","what":"What each inference provider declares to OpenRouter for an open-weight model (context length, maximum output, quantization, accepted parameters), next to the official reference config. Declared values; no request is sent to any model.","headline":"24 of the 82 provider endpoints serving 6 open-weight models declare at least one material setting that differs from the official reference (29.3%); 26 of the 71 whose quantization could be compared declare none, so it cannot be checked.","stats":{"models":6,"endpoints":82,"withDifference":32,"withMaterial":24,"quantUndisclosed":26,"quantComparable":71,"byKind":{"context":17,"quantization":5,"tools":16,"reasoning":0,"output-cap":5},"lastRead":"2026-10-19T19:49:59.892Z","firstRead":"2026-10-19T19:49:59.892Z"},"method":["Once a day the tracker reads OpenRouter's public endpoint list for each model (one row per provider deployment: declared context length, declared maximum output, declared quantization, and the request parameters it accepts) and the official repository's config.json at its head commit. Both reads are anonymous. A new row is recorded only when either moved, so the history below is a list of real changes.","Every value in the tables is what a provider declares to OpenRouter. No request is sent to any model, so nothing here says how an endpoint behaves: one that declares a 131,072-token context may still cut off earlier, and one that declares the official quantization may still serve something else.","A quantization the provider does not declare cannot be compared, and is counted as undisclosed rather than as matching. A quantization of higher precision than the official weights is shown and not counted as a difference.","Material means one of: a declared context under 95% of the reference config, a declared quantization other than the official release, no tools parameter on a model that supports tool calling, no reasoning control on a model that exposes one, or a reasoning model capped under 8,192 output tokens. A context between 95% and 100% of the reference is listed and not counted as material.","Sampling defaults, chat templates and stop tokens are not in the endpoint metadata and are not compared here. Moonshot's K2 Vendor Verifier (github.com/MoonshotAI/K2-Vendor-Verifier) measures tool-call behaviour across providers for Kimi K2 by sending requests; this page does not.","OpenRouter lists only the providers that serve through it. Nothing here ranks providers, and a provider with no listed difference has not been shown to match the reference, only to declare nothing that differs."],"rawData":"data/provider-conformance/snapshots.jsonl and runs.jsonl in the Neurarch repository.","models":[{"slug":"openai/gpt-oss-120b","page":"https://www.neurarch.com/providers/openai/gpt-oss-120b.html","reference":{"repo":"openai/gpt-oss-120b","sha":"b5c939de8f754692c1647ca79fbf85e8c1e70f8a","context":131072,"config":"https://huggingface.co/openai/gpt-oss-120b/blob/b5c939de8f754692c1647ca79fbf85e8c1e70f8a/config.json","officialQuant":"MXFP4 on the MoE experts, bf16 elsewhere","quantComparable":true,"note":null,"quantNote":null},"source":"https://openrouter.ai/api/v1/models/openai/gpt-oss-120b/endpoints","readAt":"2026-10-19T19:49:59.892Z","endpoints":[{"provider":"AkashML","tag":"akashml/bf16","quant":"bf16","context":131072,"maxOut":117964,"tools":true,"responseFormat":true,"reasoning":true,"status":0,"quantState":"higher","differences":[]},{"provider":"Amazon Bedrock","tag":"amazon-bedrock","quant":null,"context":131072,"maxOut":117964,"tools":false,"responseFormat":false,"reasoning":true,"status":0,"quantState":"undisclosed","differences":[{"kind":"tools","material":true,"text":"declares no tools parameter; the official model supports tool calling"}]},{"provider":"Amazon Bedrock","tag":"amazon-bedrock/eu-west-1","quant":null,"context":131072,"maxOut":117964,"tools":false,"responseFormat":false,"reasoning":true,"status":0,"quantState":"undisclosed","differences":[{"kind":"tools","material":true,"text":"declares no tools parameter; the official model supports tool calling"}]},{"provider":"BaseTen","tag":"baseten/fp4","quant":"fp4","context":128072,"maxOut":115264,"tools":true,"responseFormat":true,"reasoning":true,"status":0,"quantState":"same","differences":[{"kind":"context","material":false,"text":"declares a 128,072-token context; the reference config says 131,072 (98%)"}]},{"provider":"BaseTen","tag":"baseten/fp4/us","quant":"fp4","context":128072,"maxOut":115264,"tools":true,"responseFormat":true,"reasoning":true,"status":0,"quantState":"same","differences":[{"kind":"context","material":false,"text":"declares a 128,072-token context; the reference config says 131,072 (98%)"}]},{"provider":"Cerebras","tag":"cerebras/fp16","quant":"fp16","context":131072,"maxOut":40960,"tools":true,"responseFormat":true,"reasoning":true,"status":0,"quantState":"higher","differences":[]},{"provider":"CoreWeave","tag":"coreweave/fp4","quant":"fp4","context":131072,"maxOut":117964,"tools":true,"responseFormat":true,"reasoning":true,"status":0,"quantState":"same","differences":[]},{"provider":"Crusoe","tag":"crusoe/bf16","quant":"bf16","context":131072,"maxOut":117964,"tools":true,"responseFormat":true,"reasoning":true,"status":0,"quantState":"higher","differences":[]},{"provider":"DeepInfra","tag":"deepinfra/bf16","quant":"bf16","context":131072,"maxOut":117964,"tools":true,"responseFormat":true,"reasoning":true,"status":0,"quantState":"higher","differences":[]},{"provider":"DekaLLM","tag":"dekallm/bf16","quant":"bf16","context":131072,"maxOut":117964,"tools":true,"responseFormat":true,"reasoning":true,"status":0,"quantState":"higher","differences":[]},{"provider":"DigitalOcean","tag":"digitalocean","quant":null,"context":128000,"maxOut":4096,"tools":false,"responseFormat":false,"reasoning":true,"status":0,"quantState":"undisclosed","differences":[{"kind":"context","material":false,"text":"declares a 128,000-token context; the reference config says 131,072 (98%)"},{"kind":"tools","material":true,"text":"declares no tools parameter; the official model supports tool calling"},{"kind":"output-cap","material":true,"text":"declares at most 4,096 output tokens, which the reasoning tokens also count against"}]},{"provider":"Google","tag":"google-vertex/global","quant":null,"context":131072,"maxOut":117964,"tools":false,"responseFormat":true,"reasoning":true,"status":-2,"quantState":"undisclosed","differences":[{"kind":"tools","material":true,"text":"declares no tools parameter; the official model supports tool calling"}]},{"provider":"Groq","tag":"groq","quant":null,"context":131072,"maxOut":65536,"tools":true,"responseFormat":true,"reasoning":true,"status":0,"quantState":"undisclosed","differences":[]},{"provider":"Mancer 2","tag":"mancer/fp8","quant":"fp8","context":131072,"maxOut":117964,"tools":true,"responseFormat":true,"reasoning":true,"status":0,"quantState":"higher","differences":[]},{"provider":"Mara","tag":"mara","quant":null,"context":131072,"maxOut":117964,"tools":true,"responseFormat":true,"reasoning":true,"status":-2,"quantState":"undisclosed","differences":[]},{"provider":"Near AI","tag":"near-ai/fp4","quant":"fp4","context":131072,"maxOut":117964,"tools":true,"responseFormat":true,"reasoning":true,"status":0,"quantState":"same","differences":[]},{"provider":"Nebius","tag":"nebius/fp4","quant":"fp4","context":131072,"maxOut":117964,"tools":true,"responseFormat":true,"reasoning":true,"status":0,"quantState":"same","differences":[]},{"provider":"Novita","tag":"novita/fp4","quant":"fp4","context":131072,"maxOut":32768,"tools":true,"responseFormat":true,"reasoning":true,"status":-2,"quantState":"same","differences":[]},{"provider":"Parasail","tag":"parasail/fp4","quant":"fp4","context":131072,"maxOut":117964,"tools":true,"responseFormat":true,"reasoning":true,"status":0,"quantState":"same","differences":[]},{"provider":"Phala","tag":"phala","quant":null,"context":131072,"maxOut":32768,"tools":true,"responseFormat":true,"reasoning":true,"status":-2,"quantState":"undisclosed","differences":[]},{"provider":"SambaNova","tag":"sambanova","quant":null,"context":131072,"maxOut":117964,"tools":true,"responseFormat":false,"reasoning":true,"status":0,"quantState":"undisclosed","differences":[]},{"provider":"SiliconFlow","tag":"siliconflow/fp8","quant":"fp8","context":131072,"maxOut":8192,"tools":false,"responseFormat":true,"reasoning":true,"status":0,"quantState":"higher","differences":[{"kind":"tools","material":true,"text":"declares no tools parameter; the official model supports tool calling"}]},{"provider":"Together","tag":"together","quant":null,"context":131072,"maxOut":117964,"tools":true,"responseFormat":true,"reasoning":true,"status":0,"quantState":"undisclosed","differences":[]},{"provider":"Venice","tag":"venice/int4","quant":"int4","context":128000,"maxOut":32768,"tools":true,"responseFormat":false,"reasoning":true,"status":0,"quantState":"different","differences":[{"kind":"context","material":false,"text":"declares a 128,000-token context; the reference config says 131,072 (98%)"},{"kind":"quantization","material":true,"text":"declares int4 weights; the official release is MXFP4 on the MoE experts, bf16 elsewhere"}]}]},{"slug":"openai/gpt-oss-20b","page":"https://www.neurarch.com/providers/openai/gpt-oss-20b.html","reference":{"repo":"openai/gpt-oss-20b","sha":"6cee5e81ee83917806bbde320786a8fb61efebee","context":131072,"config":"https://huggingface.co/openai/gpt-oss-20b/blob/6cee5e81ee83917806bbde320786a8fb61efebee/config.json","officialQuant":"MXFP4 on the MoE experts, bf16 elsewhere","quantComparable":true,"note":null,"quantNote":null},"source":"https://openrouter.ai/api/v1/models/openai/gpt-oss-20b/endpoints","readAt":"2026-10-19T19:49:59.892Z","endpoints":[{"provider":"AkashML","tag":"akashml/fp4","quant":"fp4","context":131072,"maxOut":117964,"tools":true,"responseFormat":true,"reasoning":true,"status":0,"quantState":"same","differences":[]},{"provider":"Amazon Bedrock","tag":"amazon-bedrock","quant":null,"context":131072,"maxOut":117964,"tools":true,"responseFormat":false,"reasoning":true,"status":0,"quantState":"undisclosed","differences":[]},{"provider":"Amazon Bedrock","tag":"amazon-bedrock/eu-west-1","quant":null,"context":131072,"maxOut":117964,"tools":true,"responseFormat":false,"reasoning":true,"status":0,"quantState":"undisclosed","differences":[]},{"provider":"CoreWeave","tag":"coreweave/fp4","quant":"fp4","context":131072,"maxOut":117964,"tools":true,"responseFormat":true,"reasoning":true,"status":0,"quantState":"same","differences":[]},{"provider":"Darkbloom","tag":"darkbloom/fp8","quant":"fp8","context":131072,"maxOut":32768,"tools":true,"responseFormat":true,"reasoning":true,"status":0,"quantState":"higher","differences":[]},{"provider":"DeepInfra","tag":"deepinfra/bf16","quant":"bf16","context":131072,"maxOut":117964,"tools":true,"responseFormat":true,"reasoning":true,"status":0,"quantState":"higher","differences":[]},{"provider":"DekaLLM","tag":"dekallm/bf16","quant":"bf16","context":131072,"maxOut":117964,"tools":true,"responseFormat":true,"reasoning":true,"status":0,"quantState":"higher","differences":[]},{"provider":"Google","tag":"google-vertex/us-central1","quant":null,"context":131072,"maxOut":32768,"tools":false,"responseFormat":true,"reasoning":true,"status":0,"quantState":"undisclosed","differences":[{"kind":"tools","material":true,"text":"declares no tools parameter; the official model supports tool calling"}]},{"provider":"Groq","tag":"groq","quant":null,"context":131072,"maxOut":65536,"tools":true,"responseFormat":true,"reasoning":true,"status":0,"quantState":"undisclosed","differences":[]},{"provider":"Parasail","tag":"parasail/fp4","quant":"fp4","context":131072,"maxOut":117964,"tools":true,"responseFormat":true,"reasoning":true,"status":0,"quantState":"same","differences":[]},{"provider":"SiliconFlow","tag":"siliconflow/fp8","quant":"fp8","context":131072,"maxOut":8192,"tools":false,"responseFormat":true,"reasoning":true,"status":0,"quantState":"higher","differences":[{"kind":"tools","material":true,"text":"declares no tools parameter; the official model supports tool calling"}]}]},{"slug":"moonshotai/kimi-k2.6","page":"https://www.neurarch.com/providers/moonshotai/kimi-k2.6.html","reference":{"repo":"moonshotai/Kimi-K2.6","sha":"7eb5002f6aadc958aed6a9177b7ed26bb94011bb","context":262144,"config":"https://huggingface.co/moonshotai/Kimi-K2.6/blob/7eb5002f6aadc958aed6a9177b7ed26bb94011bb/config.json","officialQuant":"native INT4 (group size 32)","quantComparable":true,"note":null,"quantNote":null},"source":"https://openrouter.ai/api/v1/models/moonshotai/kimi-k2.6/endpoints","readAt":"2026-10-19T19:49:59.892Z","endpoints":[{"provider":"AtlasCloud","tag":"atlas-cloud/int4","quant":"int4","context":262144,"maxOut":235929,"tools":true,"responseFormat":true,"reasoning":true,"status":0,"quantState":"same","differences":[]},{"provider":"Baidu","tag":"baidu/fp4","quant":"fp4","context":262144,"maxOut":235929,"tools":true,"responseFormat":true,"reasoning":true,"status":0,"quantState":"different","differences":[{"kind":"quantization","material":true,"text":"declares fp4 weights; the official release is native INT4 (group size 32)"}]},{"provider":"Chutes","tag":"chutes/int4","quant":"int4","context":262144,"maxOut":65535,"tools":true,"responseFormat":false,"reasoning":true,"status":0,"quantState":"same","differences":[]},{"provider":"Cloudflare","tag":"cloudflare","quant":null,"context":262144,"maxOut":235929,"tools":true,"responseFormat":true,"reasoning":true,"status":0,"quantState":"undisclosed","differences":[]},{"provider":"CoreWeave","tag":"coreweave/fp4","quant":"fp4","context":262144,"maxOut":235929,"tools":true,"responseFormat":true,"reasoning":true,"status":0,"quantState":"different","differences":[{"kind":"quantization","material":true,"text":"declares fp4 weights; the official release is native INT4 (group size 32)"}]},{"provider":"Crusoe","tag":"crusoe/bf16","quant":"bf16","context":262144,"maxOut":235929,"tools":true,"responseFormat":true,"reasoning":true,"status":0,"quantState":"higher","differences":[]},{"provider":"DeepInfra","tag":"deepinfra/fp4","quant":"fp4","context":262144,"maxOut":16384,"tools":true,"responseFormat":true,"reasoning":true,"status":0,"quantState":"different","differences":[{"kind":"quantization","material":true,"text":"declares fp4 weights; the official release is native INT4 (group size 32)"}]},{"provider":"DigitalOcean","tag":"digitalocean","quant":null,"context":262144,"maxOut":235929,"tools":true,"responseFormat":true,"reasoning":true,"status":0,"quantState":"undisclosed","differences":[]},{"provider":"GMICloud","tag":"gmicloud/fp8","quant":"fp8","context":262144,"maxOut":235929,"tools":true,"responseFormat":true,"reasoning":true,"status":0,"quantState":"higher","differences":[]},{"provider":"Inceptron","tag":"inceptron/int4","quant":"int4","context":262144,"maxOut":235929,"tools":true,"responseFormat":true,"reasoning":true,"status":0,"quantState":"same","differences":[]},{"provider":"Moonshot AI","tag":"moonshotai/int4","quant":"int4","context":262144,"maxOut":235929,"tools":true,"responseFormat":true,"reasoning":true,"status":0,"quantState":"same","differences":[]},{"provider":"Novita","tag":"novita","quant":null,"context":262144,"maxOut":235929,"tools":true,"responseFormat":true,"reasoning":true,"status":0,"quantState":"undisclosed","differences":[]},{"provider":"Parasail","tag":"parasail/int4","quant":"int4","context":262144,"maxOut":235929,"tools":true,"responseFormat":true,"reasoning":true,"status":0,"quantState":"same","differences":[]},{"provider":"Phala","tag":"phala","quant":null,"context":262144,"maxOut":235929,"tools":true,"responseFormat":true,"reasoning":true,"status":0,"quantState":"undisclosed","differences":[]},{"provider":"SiliconFlow","tag":"siliconflow/fp8","quant":"fp8","context":262144,"maxOut":235929,"tools":true,"responseFormat":true,"reasoning":true,"status":0,"quantState":"higher","differences":[]},{"provider":"StreamLake","tag":"streamlake/fp8","quant":"fp8","context":256000,"maxOut":230400,"tools":true,"responseFormat":true,"reasoning":true,"status":0,"quantState":"higher","differences":[{"kind":"context","material":false,"text":"declares a 256,000-token context; the reference config says 262,144 (98%)"}]},{"provider":"Venice","tag":"venice/int4","quant":"int4","context":256000,"maxOut":65536,"tools":true,"responseFormat":true,"reasoning":true,"status":0,"quantState":"same","differences":[{"kind":"context","material":false,"text":"declares a 256,000-token context; the reference config says 262,144 (98%)"}]}]},{"slug":"qwen/qwen3-235b-a22b-2507","page":"https://www.neurarch.com/providers/qwen/qwen3-235b-a22b-2507.html","reference":{"repo":"Qwen/Qwen3-235B-A22B-Instruct-2507","sha":"ac9c66cc9b46af7306746a9250f23d47083d689e","context":262144,"config":"https://huggingface.co/Qwen/Qwen3-235B-A22B-Instruct-2507/blob/ac9c66cc9b46af7306746a9250f23d47083d689e/config.json","officialQuant":"bf16, plus an official FP8 release (Qwen/Qwen3-235B-A22B-Instruct-2507-FP8)","quantComparable":true,"note":null,"quantNote":null},"source":"https://openrouter.ai/api/v1/models/qwen/qwen3-235b-a22b-2507/endpoints","readAt":"2026-10-19T19:49:59.892Z","endpoints":[{"provider":"DeepInfra","tag":"deepinfra/fp8","quant":"fp8","context":262144,"maxOut":16384,"tools":true,"responseFormat":true,"reasoning":false,"status":0,"quantState":"same","differences":[]},{"provider":"GMICloud","tag":"gmicloud/fp8","quant":"fp8","context":262144,"maxOut":235929,"tools":true,"responseFormat":true,"reasoning":false,"status":0,"quantState":"same","differences":[]},{"provider":"Google","tag":"google-vertex/us-south1","quant":null,"context":262144,"maxOut":16384,"tools":true,"responseFormat":true,"reasoning":false,"status":0,"quantState":"undisclosed","differences":[]},{"provider":"Google","tag":"google-vertex/us-south1","quant":null,"context":262144,"maxOut":16384,"tools":false,"responseFormat":true,"reasoning":false,"status":0,"quantState":"undisclosed","differences":[{"kind":"tools","material":true,"text":"declares no tools parameter; the official model supports tool calling"}]},{"provider":"Nebius","tag":"nebius/fp8","quant":"fp8","context":262144,"maxOut":235929,"tools":true,"responseFormat":true,"reasoning":false,"status":0,"quantState":"same","differences":[]},{"provider":"Parasail","tag":"parasail/fp8","quant":"fp8","context":131072,"maxOut":117964,"tools":true,"responseFormat":true,"reasoning":false,"status":0,"quantState":"same","differences":[{"kind":"context","material":true,"text":"declares a 131,072-token context; the reference config says 262,144 (50%)"}]},{"provider":"Venice","tag":"venice/fp8","quant":"fp8","context":128000,"maxOut":16384,"tools":true,"responseFormat":true,"reasoning":false,"status":0,"quantState":"same","differences":[{"kind":"context","material":true,"text":"declares a 128,000-token context; the reference config says 262,144 (49%)"}]}]},{"slug":"deepseek/deepseek-v3.2","page":"https://www.neurarch.com/providers/deepseek/deepseek-v3.2.html","reference":{"repo":"deepseek-ai/DeepSeek-V3.2","sha":"a7e62ac04ecb2c0a54d736dc46601c5606cf10a6","context":163840,"config":"https://huggingface.co/deepseek-ai/DeepSeek-V3.2/blob/a7e62ac04ecb2c0a54d736dc46601c5606cf10a6/config.json","officialQuant":"FP8 e4m3, 128x128 blocks","quantComparable":true,"note":null,"quantNote":null},"source":"https://openrouter.ai/api/v1/models/deepseek/deepseek-v3.2/endpoints","readAt":"2026-10-19T19:49:59.892Z","endpoints":[{"provider":"AtlasCloud","tag":"atlas-cloud/fp8","quant":"fp8","context":163840,"maxOut":147456,"tools":true,"responseFormat":true,"reasoning":true,"status":0,"quantState":"same","differences":[]},{"provider":"Baidu","tag":"baidu/fp8","quant":"fp8","context":131072,"maxOut":65536,"tools":false,"responseFormat":true,"reasoning":true,"status":0,"quantState":"same","differences":[{"kind":"context","material":true,"text":"declares a 131,072-token context; the reference config says 163,840 (80%)"},{"kind":"tools","material":true,"text":"declares no tools parameter; the official model supports tool calling"}]},{"provider":"DeepInfra","tag":"deepinfra/fp4","quant":"fp4","context":163840,"maxOut":16384,"tools":true,"responseFormat":true,"reasoning":true,"status":0,"quantState":"different","differences":[{"kind":"quantization","material":true,"text":"declares fp4 weights; the official release is FP8 e4m3, 128x128 blocks"}]},{"provider":"DigitalOcean","tag":"digitalocean","quant":null,"context":163840,"maxOut":147456,"tools":false,"responseFormat":false,"reasoning":true,"status":0,"quantState":"undisclosed","differences":[{"kind":"tools","material":true,"text":"declares no tools parameter; the official model supports tool calling"}]},{"provider":"Friendli","tag":"friendli","quant":null,"context":163840,"maxOut":147456,"tools":true,"responseFormat":true,"reasoning":true,"status":0,"quantState":"undisclosed","differences":[]},{"provider":"GMICloud","tag":"gmicloud/fp8","quant":"fp8","context":163840,"maxOut":147456,"tools":true,"responseFormat":false,"reasoning":true,"status":0,"quantState":"same","differences":[]},{"provider":"Google","tag":"google-vertex","quant":null,"context":163840,"maxOut":65536,"tools":true,"responseFormat":true,"reasoning":true,"status":0,"quantState":"undisclosed","differences":[]},{"provider":"Mara","tag":"mara","quant":null,"context":32768,"maxOut":7168,"tools":false,"responseFormat":true,"reasoning":true,"status":0,"quantState":"undisclosed","differences":[{"kind":"context","material":true,"text":"declares a 32,768-token context; the reference config says 163,840 (20%)"},{"kind":"tools","material":true,"text":"declares no tools parameter; the official model supports tool calling"},{"kind":"output-cap","material":true,"text":"declares at most 7,168 output tokens, which the reasoning tokens also count against"}]},{"provider":"Phala","tag":"phala","quant":null,"context":163840,"maxOut":64000,"tools":true,"responseFormat":true,"reasoning":true,"status":0,"quantState":"undisclosed","differences":[]},{"provider":"SambaNova","tag":"sambanova","quant":null,"context":32768,"maxOut":7168,"tools":false,"responseFormat":false,"reasoning":true,"status":-2,"quantState":"undisclosed","differences":[{"kind":"context","material":true,"text":"declares a 32,768-token context; the reference config says 163,840 (20%)"},{"kind":"tools","material":true,"text":"declares no tools parameter; the official model supports tool calling"},{"kind":"output-cap","material":true,"text":"declares at most 7,168 output tokens, which the reasoning tokens also count against"}]},{"provider":"SiliconFlow","tag":"siliconflow/fp8","quant":"fp8","context":163840,"maxOut":147456,"tools":true,"responseFormat":true,"reasoning":true,"status":0,"quantState":"same","differences":[]},{"provider":"Venice","tag":"venice","quant":null,"context":160000,"maxOut":32768,"tools":true,"responseFormat":true,"reasoning":true,"status":0,"quantState":"undisclosed","differences":[{"kind":"context","material":false,"text":"declares a 160,000-token context; the reference config says 163,840 (98%)"}]}]},{"slug":"meta-llama/llama-3.3-70b-instruct","page":"https://www.neurarch.com/providers/meta-llama/llama-3.3-70b-instruct.html","reference":{"repo":"unsloth/Llama-3.3-70B-Instruct","sha":"99cd0d2c829e92a67c844f9144c2509632e5c87f","context":131072,"config":"https://huggingface.co/unsloth/Llama-3.3-70B-Instruct/blob/99cd0d2c829e92a67c844f9144c2509632e5c87f/config.json","officialQuant":"no publicly verifiable reference","quantComparable":false,"note":"Meta's repository (meta-llama/Llama-3.3-70B-Instruct) is gated and cannot be read anonymously, so the context length is read from the unsloth mirror of the same weights.","quantNote":"Meta publishes no reference quantization that can be read anonymously (its FP8 repository answers 401 to an anonymous request, which does not say whether it exists), so the quantization each provider declares is shown and not compared."},"source":"https://openrouter.ai/api/v1/models/meta-llama/llama-3.3-70b-instruct/endpoints","readAt":"2026-10-19T19:49:59.892Z","endpoints":[{"provider":"AkashML","tag":"akashml/fp8","quant":"fp8","context":131072,"maxOut":128000,"tools":true,"responseFormat":true,"reasoning":false,"status":0,"quantState":"not-compared","differences":[]},{"provider":"Cloudflare","tag":"cloudflare/fp8","quant":"fp8","context":24000,"maxOut":21600,"tools":false,"responseFormat":true,"reasoning":false,"status":0,"quantState":"not-compared","differences":[{"kind":"context","material":true,"text":"declares a 24,000-token context; the reference config says 131,072 (18%)"},{"kind":"tools","material":true,"text":"declares no tools parameter; the official model supports tool calling"}]},{"provider":"CoreWeave","tag":"coreweave/fp16","quant":"fp16","context":128000,"maxOut":115200,"tools":true,"responseFormat":true,"reasoning":false,"status":0,"quantState":"not-compared","differences":[{"kind":"context","material":false,"text":"declares a 128,000-token context; the reference config says 131,072 (98%)"}]},{"provider":"DeepInfra","tag":"deepinfra/turbo","quant":"fp8","context":131072,"maxOut":16384,"tools":true,"responseFormat":true,"reasoning":false,"status":0,"quantState":"not-compared","differences":[]},{"provider":"Google","tag":"google-vertex","quant":null,"context":128000,"maxOut":115200,"tools":false,"responseFormat":true,"reasoning":false,"status":0,"quantState":"not-compared","differences":[{"kind":"context","material":false,"text":"declares a 128,000-token context; the reference config says 131,072 (98%)"},{"kind":"tools","material":true,"text":"declares no tools parameter; the official model supports tool calling"}]},{"provider":"Google","tag":"google-vertex/us-central1","quant":null,"context":128000,"maxOut":8192,"tools":true,"responseFormat":true,"reasoning":false,"status":0,"quantState":"not-compared","differences":[{"kind":"context","material":false,"text":"declares a 128,000-token context; the reference config says 131,072 (98%)"}]},{"provider":"Groq","tag":"groq","quant":null,"context":131072,"maxOut":32768,"tools":true,"responseFormat":true,"reasoning":false,"status":0,"quantState":"not-compared","differences":[]},{"provider":"Novita","tag":"novita/bf16","quant":"bf16","context":12288,"maxOut":11059,"tools":true,"responseFormat":false,"reasoning":false,"status":0,"quantState":"not-compared","differences":[{"kind":"context","material":true,"text":"declares a 12,288-token context; the reference config says 131,072 (9%)"}]},{"provider":"Parasail","tag":"parasail/fp8","quant":"fp8","context":131072,"maxOut":16384,"tools":false,"responseFormat":true,"reasoning":false,"status":0,"quantState":"not-compared","differences":[{"kind":"tools","material":true,"text":"declares no tools parameter; the official model supports tool calling"}]},{"provider":"SambaNova","tag":"sambanova-turbo","quant":null,"context":131072,"maxOut":3072,"tools":false,"responseFormat":true,"reasoning":false,"status":0,"quantState":"not-compared","differences":[{"kind":"tools","material":true,"text":"declares no tools parameter; the official model supports tool calling"},{"kind":"output-cap","material":false,"text":"declares at most 3,072 output tokens"}]},{"provider":"Together","tag":"together","quant":null,"context":131072,"maxOut":2048,"tools":true,"responseFormat":true,"reasoning":false,"status":0,"quantState":"not-compared","differences":[{"kind":"output-cap","material":false,"text":"declares at most 2,048 output tokens"}]}]}],"history":[]}