SF·SIX / SOURCE-BACKED RESEARCH
AI API pricing, live
Live API prices for the current flagship, small and open-weight language models and for image, video and speech generation, ranked from the providers SF·SIX tracks.
Of 18 flagship models tracked, DeepSeek V4 Pro 0813 has the lowest standard price at $0.1333 per 1M input and $0.40 per 1M output tokens; GPT-6 Astra Pro is the most expensive at $10.00 / $50.00. Prices are USD per 1M tokens at the standard service tier in the default region: the cheapest qualifying provider per model, with flex, batch, priority and fast endpoints and unusually low outliers never used as the headline. “Blended” weights input and output 3:1.
Flagship LLMs
| Model | Input / 1M | Output / 1M | Blended 3:1 | Cheapest standard provider | Context (tokens) | OpenRouter tokens / day | AA intelligence |
|---|---|---|---|---|---|---|---|
| DeepSeek V4 Pro 0813DeepSeek | $0.1333 | $0.40 | $0.20 | Baidufp8 | 1,048,576 | 161.08B | 36.0 |
| MiniMax M3MiniMax | $0.23 | $0.96 | $0.4125 | CoreWeavefp4 | 1,048,576 | 210.13B | 29.2 |
| MiMo-V2.6-ProXiaomi | $0.43 | $0.87 | $0.54 | DeepInfrafp8 | 1,050,000 | 209.34B | 46.3 |
| GLM 5.3Z.ai (Zhipu) | $0.50 | $1.51 | $0.7513 | Sail Researchfp8 | 1,048,576 | 942.62B | 44.8 |
| Mistral Large 4Mistral AI | $0.68 | $2.09 | $1.03 | Mistral AI | 1,048,576 | Not in named rows | 38.4 |
| Step 5 PreviewStepFun | $1.00 | $2.70 | $1.43 | StepFunfp8 | 1,048,576 | 179.04B | 43.7 |
| Muse Spark 1.3Meta | $1.25 | $4.25 | $2.00 | Meta | 1,048,576 | Not in named rows | 48.1 |
| Kimi K3Moonshot AI | $0.504 | $8.10 | $2.40 | Sail Researchfp4 | 1,048,576 | 203.41B | 43.6 |
| Grok 4.7xAI | $2.00 | $6.00 | $3.00 | xAI | 500,000 | Not in named rows | 46.4 |
| Qwen3.8 Max (0902)Qwen (Alibaba) | $2.00 | $6.00 | $3.00 | Alibaba | 1,000,000 | Not in named rows | 45.4 |
| Claude Sonnet 5.5Anthropic | $2.00 | $10.00 | $4.00 | Anthropic | 1,000,000 | 396.5B | 56.0 |
| GPT-6.1 SolOpenAI | $2.00 | $10.00 | $4.00 | OpenAI | 1,050,000 | 293.83B | 51.8 |
| GPT-6.1 Sol ProOpenAI | $2.00 | $10.00 | $4.00 | OpenAI | 1,050,000 | Not in named rows | Not reported |
| Claude Opus 5.5Anthropic | $4.00 | $20.00 | $8.00 | Anthropic | 1,000,000 | 933.37B | 57.6 |
| Muse WonderMeta | $4.00 | $20.00 | $8.00 | Meta | 1,048,576 | Not in named rows | Not reported |
| Claude Fable 5.1Anthropic | $10.00 | $50.00 | $20.00 | Anthropic | 1,000,000 | Not in named rows | 53.4 |
| GPT-6 AstraOpenAI | $10.00 | $50.00 | $20.00 | OpenAI | 1,050,000 | 61.76B | 52.7 |
| GPT-6 Astra ProOpenAI | $10.00 | $50.00 | $20.00 | OpenAI | 1,050,000 | Not in named rows | Not reported |
Each lab’s current top models, chosen by the SF·SIX research desk. Cheapest first.
Small and fast LLMs
| Model | Input / 1M | Output / 1M | Blended 3:1 | Cheapest standard provider | Context (tokens) | OpenRouter tokens / day | AA intelligence |
|---|---|---|---|---|---|---|---|
| DeepSeek V4 Flash 0423DeepSeek | $0.0181 | $0.0361 | $0.0226 | Baidufp8 | 1,048,576 | 423.23B | 24.4 |
| Ling 3.0 FlashinclusionAI | $0.021 | $0.063 | $0.0315 | Novita AI | 262,144 | Not in named rows | Not reported |
| Nex-N2.5-MiniNex AGI | $0.025 | $0.10 | $0.0438 | Nex AGIbf16 | 262,144 | Not in named rows | Not reported |
| Qwen3.7 FlashQwen (Alibaba) | $0.03 | $0.13 | $0.055 | Alibaba | 1,000,000 | Not in named rows | Not reported |
| Solar Mini 4Upstage | $0.05 | $0.20 | $0.0875 | Upstage | 524,288 | 146.49B | 24.1 |
| DeepSeek V4 Flash 0731DeepSeek | $0.0598 | $0.1795 | $0.0898 | Baidufp8 | 1,048,576 | 620.6B | 34.3 |
| Ling 3.0 Flash VLinclusionAI | $0.06 | $0.18 | $0.09 | DeepInfrafp16 | 131,072 | Not in named rows | 24.6 |
| Schematron V2 SmallInference.net | $0.05 | $0.23 | $0.095 | InferenceNetbf16 | 128,000 | Not in named rows | Not reported |
| Ling 3.1 FlashinclusionAI | $0.075 | $0.225 | $0.1125 | Novita AI | 262,144 | Not in named rows | 41.1 |
| DeepSeek V4.1 FlashDeepSeek | $0.09 | $0.18 | $0.1125 | Decartfp4 | 1,048,576 | 5.41T | 39.5 |
| GLM 5.3 FlashZ.ai (Zhipu) | $0.075 | $0.25 | $0.1188 | DeepInfrafp4 | 1,048,576 | 1.73T | 41.8 |
| MiMo-V2.6-FlashXiaomi | $0.10 | $0.28 | $0.145 | Darkbloomfp4 | 1,050,000 | 1.44T | 37.9 |
| Claude Haiku 5.5Anthropic | $0.10 | $0.50 | $0.20 | Anthropic | 1,000,000 | 302.61B | 43.4 |
| GPT-6 LunaOpenAI | $0.10 | $0.50 | $0.20 | OpenAI | 1,050,000 | 988.36B | 38.1 |
| DeepSeek V4.1 Flash 1017DeepSeek | $0.10 | $0.50 | $0.20 | InferenceNetfp8 | 1,048,576 | Not in named rows | Not reported |
| GPT-6 Luna ProOpenAI | $0.10 | $0.50 | $0.20 | OpenAI | 1,050,000 | 62.48B | Not reported |
| Qwen3.8 FlashQwen (Alibaba) | $0.15 | $0.47 | $0.23 | Alibaba | 1,000,000 | 81.75B | Not reported |
| Qwen3.8 Omni FlashQwen (Alibaba) | $0.15 | $0.47 | $0.23 | OpenRouter (auto-routed) | 1,000,000 | Not in named rows | Not reported |
| DeepSeek V4 Flash Vision ExpDeepSeek | $0.2156 | $0.6468 | $0.3234 | DeepInfrafp8 | 1,048,576 | Not in named rows | 34.8 |
| Qwen3.6 FlashQwen (Alibaba) | $0.1875 | $1.13 | $0.4219 | Alibaba | 1,000,000 | Not in named rows | Not reported |
Small and fast tiers (Haiku, Flash, Lite, Mini, Nano, Luna and similar) released in the last six months, cheapest first. Top 20.
Open-weight LLMs
| Model | Input / 1M | Output / 1M | Blended 3:1 | Cheapest standard provider | Context (tokens) | OpenRouter tokens / day | Weights |
|---|---|---|---|---|---|---|---|
| GLM 5.3Z.ai (Zhipu) | $0.50 | $1.51 | $0.7513 | Sail Researchfp8 | 1,048,576 | 942.62B | zai-org/GLM-5.3 |
| DeepSeek V4.1 FlashDeepSeek | $0.09 | $0.18 | $0.1125 | Decartfp4 | 1,048,576 | 5.41T | deepseek-ai/DeepSeek-V4.1-Flash |
| GLM 5.3 FlashZ.ai (Zhipu) | $0.075 | $0.25 | $0.1188 | DeepInfrafp4 | 1,048,576 | 1.73T | zai-org/GLM-5.3-Flash |
| GLM 5.2Z.ai (Zhipu) | $0.189 | $0.594 | $0.2903 | Baidufp8 | 1,048,576 | 140.87B | zai-org/GLM-5.2 |
| DeepSeek V4 Flash 0731DeepSeek | $0.0598 | $0.1795 | $0.0898 | Baidufp8 | 1,048,576 | 620.6B | deepseek-ai/DeepSeek-V4-Flash-0731 |
| gpt-oss-120bOpenAI | $0.03 | $0.15 | $0.06 | Near AIfp4 | 131,072 | 86.23B | openai/gpt-oss-120b |
| DeepSeek V4 Pro 0813DeepSeek | $0.1333 | $0.40 | $0.20 | Baidufp8 | 1,048,576 | 161.08B | deepseek-ai/DeepSeek-V4-Pro-0813 |
| Kimi K3Moonshot AI | $0.504 | $8.10 | $2.40 | Sail Researchfp4 | 1,048,576 | 203.41B | moonshotai/Kimi-K3 |
| Qwen3.8 27BQwen (Alibaba) | $0.039 | $1.25 | $0.3418 | Chutesfp8 | 1,000,000 | 62.19B | Qwen/Qwen3.8-27B |
| Kimi K2.6Moonshot AI | $0.4323 | $1.82 | $0.7792 | Baidufp4 | 262,144 | Not in named rows | moonshotai/Kimi-K2.6 |
| DeepSeek V4 Flash 0423DeepSeek | $0.0181 | $0.0361 | $0.0226 | Baidufp8 | 1,048,576 | 423.23B | deepseek-ai/DeepSeek-V4-Flash |
| DeepSeek V4 Pro 0423DeepSeek | $0.1775 | $0.3549 | $0.2218 | Baidufp8 | 1,048,576 | 85.51B | deepseek-ai/DeepSeek-V4-Pro |
| Gemma 4 26B A4BGoogle DeepMind | $0.042 | $0.22 | $0.0865 | Darkbloom | 262,144 | Not in named rows | google/gemma-4-26B-A4B-it |
| MiniMax M3MiniMax | $0.23 | $0.96 | $0.4125 | CoreWeavefp4 | 1,048,576 | 210.13B | MiniMaxAI/MiniMax-M3 |
| Gemma 4 31BGoogle DeepMind | $0.10 | $0.34 | $0.16 | CoreWeavefp4 | 262,144 | 73.13B | google/gemma-4-31B-it |
| Kimi K2.7 CodeMoonshot AI | $0.7125 | $3.00 | $1.28 | StreamLake | 262,144 | Not in named rows | moonshotai/Kimi-K2.7-Code |
| GLM 5.1Z.ai (Zhipu) | $0.9646 | $3.03 | $1.48 | Baidufp8 | 204,800 | Not in named rows | zai-org/GLM-5.1 |
| Llama 3.3 70B InstructMeta | $0.135 | $0.40 | $0.2013 | Novita AIbf16 | 131,072 | Not in named rows | meta-llama/Llama-3.3-70B-Instruct |
| Qwen3.6 35B A3BQwen (Alibaba) | $0.05 | $0.70 | $0.2125 | Darkbloomfp4 | 262,144 | 386.14B | Qwen/Qwen3.6-35B-A3B |
| gpt-oss-20bOpenAI | $0.018 | $0.09 | $0.036 | Darkbloomfp8 | 131,072 | Not in named rows | openai/gpt-oss-20b |
Models whose weights are published on the Hugging Face Hub (the repository of the same name that Hugging Face Inference Providers serve), most widely served first. Top 20.
Image generation: cheapest per image
| Model | Price per image | Provider | Observed |
|---|---|---|---|
| Luma Uni-1 Text to ImageLuma AI | $0.003 | fal.ai | 2026-10-17 |
| Luma Uni-1 Text to Image MaxLuma AI | $0.003 | fal.ai | 2026-10-17 |
| VecglypherUnattributed (via fal) | $0.005 | fal.ai | 2026-10-17 |
| Recraft V4.1 Flash Text to ImageRecraft | $0.007 | fal.ai | 2026-10-17 |
| Meta Muse Image Text to ImageMeta | $0.01 | fal.ai | 2026-10-17 |
| MiniMax (Hailuo AI) Text to ImageMiniMax | $0.01 | fal.ai | 2026-10-17 |
| Wan (v2.2-5b / text-to-image)Qwen (Alibaba) | $0.016 | fal.ai | 2026-10-17 |
| Grok Imagine ImagexAI | $0.02 | fal.ai | 2026-10-17 |
| Ideogram V2A TurboIdeogram | $0.025 | fal.ai | 2026-10-17 |
| Wan (v2.2-a14b / text-to-image)Qwen (Alibaba) | $0.025 | fal.ai | 2026-10-17 |
Video generation: cheapest per second of video
| Model | Price per second | Provider | Observed |
|---|---|---|---|
| PixVerse C1 Text To VideoPixVerse | $0.005 | fal.ai | 2026-10-17 |
| Grok Imagine Video 1.5 Text to VideoxAI | $0.01 | fal.ai | 2026-10-17 |
| H3 Max Turbo Text to VideoMiniMax | $0.0163 | fal.ai | 2026-10-17 |
| LTX-Video 13B 0.9.8 DistilledLightricks | $0.02 | fal.ai | 2026-10-17 |
| Flux 3 Text To Video DraftBlack Forest Labs | $0.03 | fal.ai | 2026-10-17 |
| Gemini Omni Flash 1.1 Text to VideoGoogle DeepMind | $0.03 | fal.ai | 2026-10-17 |
| Luma Ray 3.2 Text to VideoLuma AI | $0.03 | fal.ai | 2026-10-17 |
| MiniMax H3 Max Text to VideoMiniMax | $0.0325 | fal.ai | 2026-10-17 |
| Heygen Video AgentHeyGen | $0.034 | fal.ai | 2026-10-17 |
| MiniMax Hailuo 02 [Standard] (Text to Video)MiniMax | $0.045 | fal.ai | 2026-10-17 |
Text to speech: cheapest per 1M characters
| Model | Price per 1M characters | Provider | Observed |
|---|---|---|---|
| Async Text to Speech Pro V1.0async | $10.00 | fal.ai | 2026-10-17 |
| Kokoro TTSKokoro · 9 variants at this price | $20.00 | fal.ai | 2026-10-17 |
| Chatterbox (text-to-speech / multilingual)Resemble AI · 2 variants at this price | $25.00 | fal.ai | 2026-10-17 |
| Bytedance Seed Speech Text to SpeechByteDance Seed | $30.00 | fal.ai | 2026-10-17 |
| CSM-1BUnattributed (via fal) | $30.00 | fal.ai | 2026-10-17 |
| Gemini 3.8 Flash Lite TTSGoogle DeepMind | $30.00 | fal.ai | 2026-10-17 |
| Chatterboxhd (text-to-speech)Resemble AI | $40.00 | fal.ai | 2026-10-17 |
| DiaUnattributed (via fal) | $40.00 | fal.ai | 2026-10-17 |
| Eleven v4 TurboElevenLabs | $40.00 | fal.ai | 2026-10-17 |
| Gemini 3.8 Flash TTSGoogle DeepMind | $45.00 | fal.ai | 2026-10-17 |
API pricing by developer and provider
By developer
By inference provider
Every model of one organisation in one table. What moved recently: AI API price changes · New AI models.
How these prices are compiled
Prices are USD per 1M tokens at the standard service tier in the default region: the cheapest qualifying provider per model, with flex, batch, priority and fast endpoints and unusually low outliers never used as the headline. “Blended” weights input and output 3:1. OpenRouter tokens per day are the latest full UTC day of OpenRouter’s public rankings (named rows only). AA intelligence is the Artificial Analysis Intelligence Index reported through OpenRouter (CC BY 4.0); “Not reported” means no score is available. Missing prices are unknown, not free.