Providers

• Voxtral-Mini-3B-2507×0.2
  $0.04 / $0.04×0.8 / ×0.1
• Gemma-3-4B-IT×0.3
  $0.04 / $0.08 · Arena rank #264×0.8 / ×0.2
• Llama3-2-1B×0.6
  $0.10 / $0.10×2.0 / ×0.3
• Ministral-3-3b-Instruct×0.6
  $0.10 / $0.10×2.0 / ×0.3
• GPT-OSS-Safeguard-20B×0.7
  $0.07 / $0.20×1.4 / ×0.5
• Nemotron-Nano-9B-V2×0.7
  $0.06 / $0.23×1.2 / ×0.6
• Nemotron-Nano-3-30B×0.7
  $0.06 / $0.24×1.2 / ×0.6
• gpt-oss-20b×0.9
  $0.07 / $0.30 · Arena rank #246×1.4 / ×0.8
• Llama3-2-3B×0.9
  $0.15 / $0.15×3.0 / ×0.4
• Ministral-3-8b-Instruct×0.9
  $0.15 / $0.15×3.0 / ×0.4

Provider fee: Prices are the US East (N. Virginia) on-demand rates from the AWS Price List. Other AWS regions are priced differently, and provisioned throughput is bought by the hour rather than by the token. Official source

Website · Pricing source

• gpt-oss-120b×2.9
  $0.35 / $0.75 · Arena rank #195×7.0 / ×1.9
• gemma-4-31b×6.9
  $0.99 / $1.49 · Arena rank #62×20 / ×3.7
• zai-glm-4.7×15
  $2.25 / $2.75×45 / ×6.9

Provider fee: Public hosted language models and exact structured per-token prices from the official public catalog. Official source

Website · Pricing source

• @cf/ibm-granite/granite-4.0-h-micro×0.3
  $0.017 / $0.112×0.3 / ×0.3
• @cf/meta/llama-3.2-1b-instruct×0.5
  $0.027 / $0.201 · Arena rank #375×0.5 / ×0.5
• @cf/mistral/mistral-7b-instruct-v0.1×0.8
  $0.11 / $0.19×2.2 / ×0.5
• @cf/meta/llama-3.2-3b-instruct×0.9
  $0.051 / $0.335 · Arena rank #346×1.0 / ×0.8
• @cf/qwen/qwen3-30b-a3b-fp8×0.9
  $0.051 / $0.335×1.0 / ×0.8
• @cf/meta/llama-3.1-8b-instruct-fp8-fast×0.9
  $0.045 / $0.384×0.9 / ×1.0
• @cf/google/gemma-4-26b-a4b-it×1.0
  $0.10 / $0.30×2.0 / ×0.8
• @cf/meta/llama-3-8b-instruct-awq×1.0
  $0.123 / $0.266×2.5 / ×0.7
• @cf/meta/llama-3.1-8b-instruct-awq×1.0
  $0.123 / $0.266×2.5 / ×0.7
• @cf/zai-org/glm-4.7-flash×1.0
  $0.06 / $0.40×1.2 / ×1.0

Provider fee: Workers AI is billed in Neurons against a paid plan, with a daily free allocation. The per-token prices here are the dollar equivalents Cloudflare states on the same page. Official source

Website · Pricing source

• mistralai/Mistral-Nemo-Instruct-2407×0.1
  $0.019 / $0.03×0.4 / ×0.1
• meta-llama/Meta-Llama-3.1-8B-Instruct-Turbo×0.2
  $0.02 / $0.04×0.4 / ×0.1
• google/gemma-4-E4B-it×0.3
  $0.02 / $0.10×0.4 / ×0.3
• mistralai/Mistral-Small-24B-Instruct-2501×0.4
  $0.05 / $0.08 · Arena rank #285×1.0 / ×0.2
• google/gemma-3-4b-it×0.4
  $0.05 / $0.10 · Arena rank #264×1.0 / ×0.3
• google/gemma-3-12b-it×0.5
  $0.05 / $0.15 · Arena rank #213×1.0 / ×0.4
• microsoft/phi-4×0.6
  $0.07 / $0.14 · Arena rank #297×1.4 / ×0.4
• nvidia/Nemotron-3-Nano-30B-A3B×0.6
  $0.05 / $0.20×1.0 / ×0.5
• google/gemma-3-27b-it×0.6
  $0.08 / $0.16 · Arena rank #181×1.6 / ×0.4
• deepseek-ai/DeepSeek-V4-Flash-0731×0.7
  $0.08 / $0.18×1.6 / ×0.5

Website · Pricing source

• accounts/fireworks/models/nemotron-lightning-3p5-30b-a3b×0.6
  $0.05 / $0.20×1.0 / ×0.5
• accounts/fireworks/models/gpt-oss-20b×0.9
  $0.07 / $0.30 · Arena rank #246×1.4 / ×0.8
• accounts/fireworks/models/deepseek-v4-flash×1.1
  $0.14 / $0.28 · Arena rank #85×2.8 / ×0.7
• accounts/fireworks/models/deepseek-v4-flash-0731×1.1
  $0.14 / $0.28×2.8 / ×0.7
• accounts/fireworks/models/gpt-oss-120b×1.8
  $0.15 / $0.60 · Arena rank #195×3.0 / ×1.5
• accounts/fireworks/models/minimax-m3×3.6
  $0.30 / $1.20 · Arena rank #76×6.0 / ×3.0
• accounts/fireworks/models/minimax-m2p7×3.6
  $0.30 / $1.20 · Arena rank #119×6.0 / ×3.0
• accounts/fireworks/models/muse-glimmer-30b×4.4
  $0.35 / $1.50×7.0 / ×3.8
• accounts/fireworks/models/qwen3p7-plus×4.8
  $0.40 / $1.60 · Arena rank #52×8.0 / ×4.0
• accounts/fireworks/models/nemotron-3-ultra-nvfp4×7.2
  $0.60 / $2.40×12 / ×6.0

Provider fee: Only Fireworks Serverless text and vision-language models are included; dedicated and on-demand deployments are excluded. Official source

Website · Pricing source

• llama-3.1-8b-instant×0.4
  $0.05 / $0.08×1.0 / ×0.2
• openai/gpt-oss-20b×0.9
  $0.075 / $0.30 · Arena rank #246×1.5 / ×0.8
• openai/gpt-oss-120b×1.8
  $0.15 / $0.60 · Arena rank #195×3.0 / ×1.5
• llama-3.3-70b-versatile×3.9
  $0.59 / $0.79×12 / ×2.0
• qwen/qwen3.6-27b×8.4
  $0.60 / $3.00×12 / ×7.5
• minimaxai/minimax-m2.7×?
  No token price published

Provider fee: Production and Preview general text-output listings are included; Contact Sales remains unpriced. Official source

Website · Pricing source

• Qwen/Qwen2.5-Coder-3B-Instructvia Nscale×0.1
  $0.01 / $0.03×0.2 / ×0.1
• Qwen/Qwen2.5-Coder-7B-Instructvia Nscale×0.1
  $0.01 / $0.03×0.2 / ×0.1
• Qwen/Qwen3-4B-Instruct-2507via Nscale×0.1
  $0.01 / $0.03×0.2 / ×0.1
• Qwen/Qwen3-4B-Thinking-2507via Nscale×0.1
  $0.01 / $0.03×0.2 / ×0.1
• meta-llama/Llama-3.1-8B-Instructvia DeepInfra×0.2
  $0.02 / $0.05 · Arena rank #320×0.4 / ×0.1
• meta-llama/Llama-3.1-8B-Instructvia Novita×0.2
  $0.02 / $0.05 · Arena rank #320×0.4 / ×0.1
• deepseek-ai/DeepSeek-R1-Distill-Llama-8Bvia Nscale×0.3
  $0.05 / $0.05×1.0 / ×0.1
• meta-llama/Llama-3.1-8B-Instructvia Nscale×0.4
  $0.06 / $0.06 · Arena rank #320×1.2 / ×0.2
• google/gemma-3-4b-itvia DeepInfra×0.4
  $0.05 / $0.10 · Arena rank #264×1.0 / ×0.3
• zai-org/AutoGLM-Phone-9B-Multilingualvia Novita×0.4
  $0.035 / $0.138×0.7 / ×0.3

Provider fee: Hugging Face routes requests to the selected inference provider and states that it adds no markup to the provider rate. Official source

Website · Pricing source

• zai-org/autoglm-phone-9b-multilingual×?
  No token price published
• deepseek/deepseek-r1-turbo×?
  No token price published
• deepseek/deepseek-v3-turbo×?
  No token price published
• deepseek/deepseek-v3-0324×?
  No token price published
• deepseek/deepseek-v3.1×?
  No token price published
• deepseek/deepseek-v3.1-terminus×?
  No token price published
• deepseek/deepseek-v3.2×?
  No token price published
• deepseek/deepseek-v3.2-exp×?
  No token price published
• deepseek/deepseek-v4-flash×?
  No token price published
• deepseek/deepseek-v4-flash-0731×?
  No token price published

Provider fee: Public Serverless LLM listings include exact representable tier bands; unsupported complex prices remain unpriced rather than flattened. Official source

Website · Pricing source

• google/lyria-3-clip-preview×0.0
  $0 / $0×0.0 / ×0.0
• google/lyria-3-pro-preview×0.0
  $0 / $0×0.0 / ×0.0
• mistralai/mistral-nemo×0.1
  $0.019 / $0.03×0.4 / ×0.1
• mistralai/mistral-small-24b-instruct-2501×0.4
  $0.05 / $0.08 · Arena rank #285×1.0 / ×0.2
• openai/gpt-oss-20b×0.4
  $0.03 / $0.13 · Arena rank #246×0.6 / ×0.3
• qwen/qwen3.7-flash×0.4
  $0.03 / $0.13×0.6 / ×0.3
• google/gemma-3-4b-it×0.4
  $0.05 / $0.10 · Arena rank #264×1.0 / ×0.3
• cohere/command-r7b-12-2024×0.5
  $0.0375 / $0.15×0.8 / ×0.4
• openai/gpt-oss-120b×0.5
  $0.03 / $0.17 · Arena rank #195×0.6 / ×0.4
• google/gemma-3n-e4b-it×0.5
  $0.06 / $0.12 · Arena rank #243×1.2 / ×0.3

Provider fee: OpenRouter charges a fee on credit purchases. It is not part of the per-token prices shown here; see the official FAQ. Official source

Website · Pricing source

• google/gemma-3n-E4B-it×0.5
  $0.06 / $0.12 · Arena rank #243×1.2 / ×0.3
• openai/gpt-oss-20b×0.6
  $0.05 / $0.20 · Arena rank #246×1.0 / ×0.5
• deepseek-ai/DeepSeek-V4-Flash-0731×1.1
  $0.14 / $0.28×2.8 / ×0.7
• Qwen/Qwen3.5-9B×1.2
  $0.17 / $0.25×3.4 / ×0.6
• openai/gpt-oss-120b×1.8
  $0.15 / $0.60 · Arena rank #195×3.0 / ×1.5
• Qwen/Qwen2.5-7B-Instruct-Turbo×1.8
  $0.30 / $0.30×6.0 / ×0.8
• google/gemma-4-31B-it×3.5
  $0.39 / $0.97×7.8 / ×2.4
• MiniMaxAI/MiniMax-M3×3.6
  $0.30 / $1.20 · Arena rank #76×6.0 / ×3.0
• Qwen/Qwen3.7-Plus×3.8
  $0.32 / $1.28 · Arena rank #52×6.4 / ×3.2
• meta-models/Muse-Glimmer-30B×4.4
  $0.35 / $1.50×7.0 / ×3.8

Provider fee: Only the public Serverless Chat catalog is included; dedicated endpoints and other inference products are excluded. Official source

Website · Pricing source

• claude-haiku-3-5×11
  $0.80 / $4.00×16 / ×10
• claude-haiku-4-5×14
  $1.00 / $5.00×20 / ×13
• claude-sonnet-5×28
  $2.00 / $10.00 · Arena rank #47i×40 / ×25
• claude-sonnet-4×42
  $3.00 / $15.00×60 / ×38
• claude-sonnet-4-5×42
  $3.00 / $15.00×60 / ×38
• claude-sonnet-4-6×42
  $3.00 / $15.00 · Arena rank #32×60 / ×38
• claude-opus-4-5×70
  $5.00 / $25.00×100 / ×63
• claude-opus-4-6×70
  $5.00 / $25.00 · Arena rank #5×100 / ×63
• claude-opus-4-7×70
  $5.00 / $25.00 · Arena rank #6×100 / ×63
• claude-opus-4-8×70
  $5.00 / $25.00 · Arena rank #29×100 / ×63

Website · Pricing source

• c4ai-aya-expanse-32b×5.0
  $0.50 / $1.50 · Arena rank #289×10 / ×3.8
• command-r-plus-08-2024×30
  $2.50 / $10.00 · Arena rank #283×50 / ×25
• c4ai-aya-vision-32b×?
  No token price published
• command-a-03-2025×?
  No token price published
• command-a-plus-05-2026×?
  No token price published
• command-a-reasoning-08-2025×?
  No token price published
• command-a-translate-08-2025×?
  No token price published
• command-a-vision-07-2025×?
  No token price published
• command-r-08-2024×?
  No token price published
• command-r7b-12-2024×?
  No token price published

Provider fee: Only exact public production API token prices are included; Trial access, private deployment, and custom sales pricing are excluded. Official source

Website · Pricing source

• deepseek-v4-flash×2.2
  $0.22 / $0.66 · Arena rank #85×4.4 / ×1.7
• deepseek-v4-pro×6.6
  $0.66 / $1.98 · Arena rank #53×13 / ×5.0

Website · Pricing source

• gemini-2.0-flash-lite×0.9
  $0.075 / $0.30×1.5 / ×0.8
• gemini-2.0-flash×1.2
  $0.10 / $0.40×2.0 / ×1.0
• gemini-2.5-flash-lite×1.2
  $0.10 / $0.40×2.0 / ×1.0
• gemini-2.5-flash-lite-preview-09-2025×1.2
  $0.10 / $0.40×2.0 / ×1.0
• gemini-3.1-flash-lite×4.0
  $0.25 / $1.50×5.0 / ×3.8
• gemini-3.5-flash-lite×6.2
  $0.30 / $2.50 · Arena rank #55×6.0 / ×6.3
• gemini-2.5-flash×6.2
  $0.30 / $2.50 · Arena rank #131×6.0 / ×6.3
• gemini-3-flash-preview×8.0
  $0.50 / $3.00×10 / ×7.5
• gemini-3.6-flash×11
  $0.75 / $3.75 · Arena rank #20i×15 / ×9.4
• gemini-3.7-flash×11
  $0.75 / $3.75×15 / ×9.4

Provider fee: Prices are the Gemini Developer API paid-tier rates from Google AI Studio. Vertex AI sells the same models under a separate price list. Official source

Website · Pricing source

• muse-spark-1.2×14
  $1.25 / $4.25 · Arena rank #4i×25 / ×11

Website · Pricing source

• MiniMax-M2.7×3.6
  $0.30 / $1.20 · Arena rank #119×6.0 / ×3.0
• MiniMax-M3×?
  $0.30 / $1.20i · Arena rank #76×6.0 / ×3.0

Website · Pricing source

• labs-leanstral-1-5×0.0
  $0 / $0×0.0 / ×0.0
• ministral-3b-2512×0.6
  $0.10 / $0.10×2.0 / ×0.3
• ministral-8b-2512×0.9
  $0.15 / $0.15×3.0 / ×0.4
• ministral-14b-2512×1.2
  $0.20 / $0.20×4.0 / ×0.5
• mistral-small-2603×1.8
  $0.15 / $0.60×3.0 / ×1.5
• codestral-2508×3.0
  $0.30 / $0.90×6.0 / ×2.3
• mistral-large-2512×5.0
  $0.50 / $1.50×10 / ×3.8
• mistral-medium-3-5×21
  $1.50 / $7.50 · Arena rank #98×30 / ×19

Website · Pricing source

• kimi-k2.6×12
  $0.95 / $4.00 · Arena rank #48×19 / ×10
• kimi-k2.7-code×12
  $0.95 / $4.00×19 / ×10
• kimi-k3×42
  $3.00 / $15.00 · Arena rank #10i×60 / ×38

Website · Pricing source

• gpt-5-nano×1.0
  $0.05 / $0.40 · Arena rank #215i×1.0 / ×1.0
• gpt-4.1-nano×1.2
  $0.10 / $0.40×2.0 / ×1.0
• gpt-4o-mini×1.8
  $0.15 / $0.60×3.0 / ×1.5
• babbage-002×2.4
  $0.40 / $0.40×8.0 / ×1.0
• gpt-5.6-luna×3.2
  $0.20 / $1.20 · Arena rank #63i×4.0 / ×3.0
• gpt-5.4-nano×3.3
  $0.20 / $1.25 · Arena rank #139i×4.0 / ×3.1
• gpt-4.1-mini×4.8
  $0.40 / $1.60×8.0 / ×4.0
• gpt-3.5-turbo-0125×5.0
  $0.50 / $1.50 · Arena rank #312×10 / ×3.8
• gpt-3.5-turbo×5.0
  $0.50 / $1.50×10 / ×3.8
• gpt-5-mini×5.0
  $0.25 / $2.00 · Arena rank #156i×5.0 / ×5.0

Website · Pricing source

• qwen-turbo×0.6
  $0.05 / $0.20×1.0 / ×0.5
• qwen3-32b×1.9
  $0.16 / $0.64 · Arena rank #204×3.2 / ×1.6
• qwen3-8b×2.1
  $0.18 / $0.70×3.6 / ×1.8
• qwen3-30b-a3b×2.4
  $0.20 / $0.80 · Arena rank #230×4.0 / ×2.0
• qwen3-30b-a3b-instruct-2507×2.4
  $0.20 / $0.80 · Arena rank #165×4.0 / ×2.0
• qwen3-235b-a22b-instruct-2507×2.8
  $0.23 / $0.92×4.6 / ×2.3
• qwen3-next-80b-a3b-instruct×3.0
  $0.15 / $1.20 · Arena rank #141×3.0 / ×3.0
• qwen3-next-80b-a3b-thinking×3.0
  $0.15 / $1.20 · Arena rank #178×3.0 / ×3.0
• qwen-plus-2025-01-25×4.0
  $0.40 / $1.20×8.0 / ×3.0
• qwen-plus-2025-04-28×4.0
  $0.40 / $1.20×8.0 / ×3.0

Provider fee: Only Qwen-authored models in the Alibaba Cloud Model Studio Singapore deployment are included. Official source

Website · Pricing source

• grok-build-0.1×8.0
  $1.00 / $2.00×20 / ×5.0
• grok-4.20-0309-non-reasoning×10
  $1.25 / $2.50×25 / ×6.3
• grok-4.20-0309-reasoning×10
  $1.25 / $2.50×25 / ×6.3
• grok-4.20-multi-agent-0309×10
  $1.25 / $2.50×25 / ×6.3
• grok-4.3×10
  $1.25 / $2.50 · Arena rank #78×25 / ×6.3
• grok-4.5×20
  $2.00 / $6.00 · Arena rank #36×40 / ×15
• grok-4.6×20
  $2.00 / $6.00 · Arena rank #46i×40 / ×15

Website · Pricing source

• mimo-v2.5×1.1
  $0.14 / $0.28 · Arena rank #90×2.8 / ×0.7
• mimo-v2.5-pro×3.5
  $0.435 / $0.87 · Arena rank #40×8.7 / ×2.2

Provider fee: Only the official Overseas Pricing token rates are included; domestic RMB pricing and hosted web-search fees are excluded. Official source

Website · Pricing source

• GLM-4.5-Flash×0.0
  $0 / $0×0.0 / ×0.0
• GLM-4.7-Flash×0.0
  $0 / $0×0.0 / ×0.0
• GLM-4-32B-0414-128K×0.6
  $0.10 / $0.10×2.0 / ×0.3
• GLM-4.7-FlashX×1.1
  $0.07 / $0.40×1.4 / ×1.0
• GLM-4.5-Air×3.0
  $0.20 / $1.10 · Arena rank #176×4.0 / ×2.8
• GLM-4.5×6.8
  $0.60 / $2.20 · Arena rank #130×12 / ×5.5
• GLM-4.6×6.8
  $0.60 / $2.20 · Arena rank #106×12 / ×5.5
• GLM-4.7×6.8
  $0.60 / $2.20×12 / ×5.5
• GLM-5×10
  $1.00 / $3.20×20 / ×8.0
• GLM-5-Turbo×13
  $1.20 / $4.00×24 / ×10

Website · Pricing source