{"ok":true,"source":"tensorfeed.ai","lastUpdated":"2026-10-26","tracked_providers":["Together AI","Fireworks","DeepInfra","Groq","OpenRouter","Replicate","Anyscale","DeepSeek","GitHub Models","Venice"],"count":10,"models":[{"modelId":"llama-4-maverick","modelName":"Llama 4 Maverick","family":"Meta","paramsB":400,"license":"Llama 4 Community License","openWeights":true,"offers":[{"provider":"OpenRouter","providerModelId":"meta-llama/llama-4-maverick","inputPrice":0.1875,"outputPrice":0.6525,"blendedPrice":0.42,"contextWindow":524288,"outputTPS":null,"features":["function-calling","vision"],"url":"https://openrouter.ai/meta-llama/llama-4-maverick","note":"Only per-token route left in this matrix; DeepInfra dropped Maverick and its model page now 404s. OpenRouter serves 524K of the model's 1M context, and EU in-region routing is not available for it"}]},{"modelId":"llama-4-scout","modelName":"Llama 4 Scout","family":"Meta","paramsB":109,"license":"Llama 4 Community License","openWeights":true,"offers":[{"provider":"DeepInfra","providerModelId":"meta-llama/Llama-4-Scout-17B-16E-Instruct","inputPrice":0.1,"outputPrice":0.3,"blendedPrice":0.2,"contextWindow":10000000,"outputTPS":170,"features":["function-calling","vision"],"url":"https://deepinfra.com/pricing","note":""},{"provider":"OpenRouter","providerModelId":"meta-llama/llama-4-scout","inputPrice":0.1,"outputPrice":0.3,"blendedPrice":0.2,"contextWindow":1310720,"outputTPS":null,"features":["function-calling","vision"],"url":"https://openrouter.ai/meta-llama/llama-4-scout","note":"Same headline price as DeepInfra, but served at 1.3M context rather than the model's full 10M"}]},{"modelId":"llama-3.1-405b","modelName":"Llama 3.1 405B Instruct","family":"Meta","paramsB":405,"license":"Llama 3.1 Community License","openWeights":true,"offers":[{"provider":"OpenRouter","providerModelId":"meta-llama/llama-3.1-405b-instruct","inputPrice":3,"outputPrice":3,"blendedPrice":3,"contextWindow":130000,"outputTPS":null,"features":["function-calling"],"url":"https://openrouter.ai/meta-llama/llama-3.1-405b-instruct","note":"Only per-token route left in this matrix; Together, Fireworks and DeepInfra dropped serverless 405B"}]},{"modelId":"llama-3.1-70b","modelName":"Llama 3.1 70B Instruct","family":"Meta","paramsB":70,"license":"Llama 3.1 Community License","openWeights":true,"offers":[{"provider":"OpenRouter","providerModelId":"meta-llama/llama-3.1-70b-instruct","inputPrice":0.4,"outputPrice":0.4,"blendedPrice":0.4,"contextWindow":130000,"outputTPS":null,"features":["function-calling"],"url":"https://openrouter.ai/meta-llama/llama-3.1-70b-instruct","note":"Only per-token route left in this matrix; Groq now serves Llama 3.3 70B in its place"}]},{"modelId":"deepseek-v4-pro","modelName":"DeepSeek V4 Pro","family":"DeepSeek","paramsB":1600,"license":"MIT","openWeights":true,"offers":[{"provider":"Together AI","providerModelId":"deepseek-ai/DeepSeek-V4-Pro","inputPrice":1.32,"outputPrice":3.96,"blendedPrice":2.64,"contextWindow":1000000,"outputTPS":90,"features":["function-calling","json-mode"],"url":"https://www.together.ai/pricing","note":"Listed as DeepSeek V4 Pro 0813; cached input is $0.13. The only large-lab serverless route left for V4 Pro"},{"provider":"OpenRouter","providerModelId":"deepseek/deepseek-v4-pro","inputPrice":0.7689,"outputPrice":1.538,"blendedPrice":1.15345,"contextWindow":1048576,"outputTPS":null,"features":["function-calling"],"url":"https://openrouter.ai/deepseek/deepseek-v4-pro","note":"Listed as the 0423 snapshot at a 56 percent discount, so the rate can move. Still the cheapest route here. The old deepseek/deepseek-chat slug tracked an alias DeepSeek retired on 2026-07-24"},{"provider":"Venice","providerModelId":"deepseek-v4-pro","inputPrice":1.65,"outputPrice":3.301,"blendedPrice":2.4755,"contextWindow":1000000,"outputTPS":null,"features":["function-calling","json-mode"],"url":"https://docs.venice.ai/overview/pricing","note":"Privacy-first host: no prompt retention. Venice prices the newer deepseek-v4-pro-0813 snapshot higher, at $1.65 in and $4.95 out"}]},{"modelId":"deepseek-v4-flash","modelName":"DeepSeek V4 Flash","family":"DeepSeek","paramsB":70,"license":"MIT","openWeights":true,"offers":[{"provider":"OpenRouter","providerModelId":"deepseek/deepseek-v4-flash","inputPrice":0.0449,"outputPrice":0.45,"blendedPrice":0.24745,"contextWindow":1000000,"outputTPS":null,"features":["function-calling"],"url":"https://openrouter.ai/deepseek/deepseek-v4-flash/pricing","note":"Input got cheaper and output got four times dearer since the last refresh; the old 60 percent promotional label is gone"},{"provider":"Venice","providerModelId":"deepseek-v4-flash","inputPrice":0.138,"outputPrice":0.275,"blendedPrice":0.2065,"contextWindow":1000000,"outputTPS":null,"features":["function-calling","json-mode"],"url":"https://docs.venice.ai/overview/pricing","note":"Listed as the 0423 snapshot. Venice also keeps a 0731 snapshot at $0.13 in and $0.26 out, and an end-to-end encrypted variant at $0.182 and $0.373"}]},{"modelId":"deepseek-v4.1-flash","modelName":"DeepSeek V4.1 Flash","family":"DeepSeek","paramsB":552,"license":"MIT","openWeights":true,"offers":[{"provider":"Together AI","providerModelId":"deepseek-ai/DeepSeek-V4.1-Flash","inputPrice":0.3,"outputPrice":1.2,"blendedPrice":0.75,"contextWindow":1000000,"outputTPS":null,"features":["function-calling","json-mode"],"url":"https://www.together.ai/pricing","note":"Cached input is $0.006, a 50x discount; the 1017 snapshot is the same price"},{"provider":"Fireworks","providerModelId":"accounts/fireworks/models/deepseek-v4p1-flash-1017","inputPrice":0.3,"outputPrice":1.2,"blendedPrice":0.75,"contextWindow":1000000,"outputTPS":null,"features":["function-calling","json-mode"],"url":"https://docs.fireworks.ai/serverless/pricing","note":"Standard tier. Priority is 25 percent more and the US-only region is 50 percent more; cached input is $0.006"},{"provider":"OpenRouter","providerModelId":"deepseek/deepseek-v4.1-flash","inputPrice":0.02,"outputPrice":1,"blendedPrice":0.51,"contextWindow":1000000,"outputTPS":null,"features":["function-calling"],"url":"https://openrouter.ai/deepseek/deepseek-v4.1-flash","note":"Cheapest input in the matrix at $0.02 per 1M"},{"provider":"Venice","providerModelId":"deepseek-v4-1-flash","inputPrice":0.3,"outputPrice":1.2,"blendedPrice":0.75,"contextWindow":1000000,"outputTPS":null,"features":["function-calling","json-mode"],"url":"https://docs.venice.ai/overview/pricing","note":"Privacy-first host: no prompt retention. An uncensored variant runs 10 percent higher at $0.33 and $1.32"}]},{"modelId":"mixtral-8x22b","modelName":"Mixtral 8x22B Instruct","family":"Mistral","paramsB":141,"license":"Apache-2.0","openWeights":true,"offers":[{"provider":"OpenRouter","providerModelId":"mistralai/mixtral-8x22b-instruct","inputPrice":2,"outputPrice":6,"blendedPrice":4,"contextWindow":65536,"outputTPS":null,"features":["function-calling"],"url":"https://openrouter.ai/mistralai/mixtral-8x22b-instruct","note":"Only per-token route left in this matrix; DeepInfra removed the model and Fireworks moved it to dedicated deployment. The price jumped from $0.65 flat to $2.00 in and $6.00 out, which makes this an expensive way to run a 2024 model"}]},{"modelId":"phi-4","modelName":"Phi-4","family":"Microsoft","paramsB":14,"license":"MIT","openWeights":true,"offers":[{"provider":"GitHub Models","providerModelId":"Phi-4","inputPrice":0,"outputPrice":0,"blendedPrice":0,"contextWindow":16384,"outputTPS":null,"features":["function-calling"],"url":"https://github.com/marketplace/models/azureml/Phi-4","note":"Free tier rate-limited (prototyping); paid via Azure AI Foundry"},{"provider":"DeepInfra","providerModelId":"microsoft/phi-4","inputPrice":0.07,"outputPrice":0.14,"blendedPrice":0.105,"contextWindow":16384,"outputTPS":220,"features":["function-calling"],"url":"https://deepinfra.com/pricing","note":""},{"provider":"OpenRouter","providerModelId":"microsoft/phi-4","inputPrice":0.07,"outputPrice":0.14,"blendedPrice":0.105,"contextWindow":16384,"outputTPS":null,"features":["function-calling"],"url":"https://openrouter.ai/microsoft/phi-4","note":""}]},{"modelId":"qwen-2.5-72b","modelName":"Qwen 2.5 72B Instruct","family":"Alibaba","paramsB":72,"license":"Qwen License","openWeights":true,"offers":[{"provider":"DeepInfra","providerModelId":"Qwen/Qwen2.5-72B-Instruct","inputPrice":0.36,"outputPrice":0.4,"blendedPrice":0.38,"contextWindow":130000,"outputTPS":80,"features":["function-calling"],"url":"https://deepinfra.com/pricing","note":""},{"provider":"OpenRouter","providerModelId":"qwen/qwen-2.5-72b-instruct","inputPrice":0.36,"outputPrice":0.4,"blendedPrice":0.38,"contextWindow":32768,"outputTPS":null,"features":["function-calling"],"url":"https://openrouter.ai/qwen/qwen-2.5-72b-instruct","note":"Now served at 33K context, down from the 130K this matrix recorded before, so long-context jobs need DeepInfra"}]}]}