# Token Pricing Table

> Neutral, real-time comparison of AI model pricing, inference costs, and token rates across OpenAI, Anthropic, Google, Meta, DeepSeek, Mistral, Qwen, xAI, Cohere, and Amazon.

## Core Resources
- [Master Pricing Table](https://tokenpricingtable.com/): Comprehensive comparative table for input tokens, output tokens, cache read/write rates, and context limits.
- [Model Data API](https://tokenpricingtable.com/models.json): Static JSON export of all 34 foundation model specifications and pricing tiers.

## Individual Model Pricing Pages
- [DeepSeek V4.1 Flash (DeepSeek)](https://tokenpricingtable.com/model/deepseek-v4-1-flash): Input 0.1500/1M, Output 0.6000/1M, 1049k context.
- [Claude Opus 5.5 (Anthropic)](https://tokenpricingtable.com/model/claude-opus-5-5): Input 4.0000/1M, Output 20.0000/1M, 1000k context.
- [GLM 5.3 Flash (Zhipu AI)](https://tokenpricingtable.com/model/glm-5-3-flash): Input 0.1500/1M, Output 0.5000/1M, 1311k context.
- [Kimi K3.1 Pro (Moonshot)](https://tokenpricingtable.com/model/kimi-k3-1-pro): Input 2.0000/1M, Output 10.0000/1M, 1049k context.
- [MiniMax M3.1 Flash (MiniMax)](https://tokenpricingtable.com/model/minimax-m3-1-flash): Input 0.1500/1M, Output 0.6000/1M, 1049k context.
- [Mercury 3 (Inception)](https://tokenpricingtable.com/model/mercury-3): Input 0.1000/1M, Output 0.3500/1M, 1000k context.
- [Step 5 Preview (StepFun)](https://tokenpricingtable.com/model/step-5-preview): Input 1.0000/1M, Output 2.7000/1M, 1000k context.
- [MiMo-V2.6 Flash (Xiaomi)](https://tokenpricingtable.com/model/mimo-v2-6-flash): Input 0.1400/1M, Output 0.2800/1M, 1050k context.
- [GPT-6 Astra (OpenAI)](https://tokenpricingtable.com/model/gpt-6-astra): Input 10.0000/1M, Output 50.0000/1M, 1050k context.
- [Claude Fable 5.1 (Anthropic)](https://tokenpricingtable.com/model/claude-fable-5-1): Input 10.0000/1M, Output 50.0000/1M, 1000k context.
- [Claude Sonnet 5.5 (Anthropic)](https://tokenpricingtable.com/model/claude-sonnet-5-5): Input 2.0000/1M, Output 10.0000/1M, 1000k context.
- [GPT-6 Luna (OpenAI)](https://tokenpricingtable.com/model/gpt-6-luna): Input 0.1000/1M, Output 0.5000/1M, 1050k context.
- [Qwen 3.8 27B (Qwen)](https://tokenpricingtable.com/model/qwen-3-8-27b): Input 0.2000/1M, Output 2.5000/1M, 1000k context.
- [GPT-6.1 Sol (OpenAI)](https://tokenpricingtable.com/model/gpt-6-1-sol): Input 2.0000/1M, Output 10.0000/1M, 1050k context.
- [Qwen 3.8 Flash (Qwen)](https://tokenpricingtable.com/model/qwen-3-8-flash): Input 0.1500/1M, Output 0.4700/1M, 1000k context.
- [Nemotron 3 Ultra (NVIDIA)](https://tokenpricingtable.com/model/nemotron-3-ultra): Input 0.6000/1M, Output 2.4000/1M, 256k context.
- [Gemini 4 Argon (Google)](https://tokenpricingtable.com/model/gemini-4-argon): Input 2.0000/1M, Output 10.0000/1M, 1049k context.
- [MiMo-V2.6 Pro (Xiaomi)](https://tokenpricingtable.com/model/mimo-v2-6-pro): Input 0.4350/1M, Output 0.8700/1M, 1050k context.
- [GLM 5.3 (Zhipu AI)](https://tokenpricingtable.com/model/glm-5-3): Input 0.0699/1M, Output 1.4000/1M, 1311k context.
- [DeepSeek V4 Pro (DeepSeek)](https://tokenpricingtable.com/model/deepseek-v4-pro): Input 0.9550/1M, Output 1.9110/1M, 1049k context.
- [Llama 4 Scout (Meta)](https://tokenpricingtable.com/model/llama-4-scout): Input 0.1000/1M, Output 0.3000/1M, 1300k context.
- [Gemini 3.8 Flash (Google)](https://tokenpricingtable.com/model/gemini-3-8-flash): Input 0.8250/1M, Output 4.1250/1M, 1049k context.
- [Grok 4.7 (xAI)](https://tokenpricingtable.com/model/grok-4-7): Input 2.0000/1M, Output 6.0000/1M, 500k context.
- [GLM 5.3 Prime (Zhipu AI)](https://tokenpricingtable.com/model/glm-5-3-prime): Input 2.8000/1M, Output 8.8000/1M, 1000k context.
- [Qwen 3.8 Max (Qwen)](https://tokenpricingtable.com/model/qwen-3-8-max): Input 2.0000/1M, Output 6.0000/1M, 1000k context.
- [Claude Sonnet 4.6 (Anthropic)](https://tokenpricingtable.com/model/claude-sonnet-4-6): Input 3.0000/1M, Output 15.0000/1M, 1000k context.
- [Gemini 3.5 Flash-Lite (Google)](https://tokenpricingtable.com/model/gemini-3-5-flash-lite): Input 0.3000/1M, Output 2.5000/1M, 1000k context.
- [Ling 3.0 Flash (InclusionAI)](https://tokenpricingtable.com/model/ling-3-0-flash): Input 0.0210/1M, Output 0.0630/1M, 262k context.
- [Llama 4 Maverick (Meta)](https://tokenpricingtable.com/model/llama-4-maverick): Input 0.1875/1M, Output 0.6525/1M, 1000k context.
- [Claude Haiku 5.5 (Anthropic)](https://tokenpricingtable.com/model/claude-haiku-5-5): Input 0.1000/1M, Output 0.5000/1M, 1000k context.
- [Mistral Large 4 (Mistral)](https://tokenpricingtable.com/model/mistral-large-4): Input 1.3600/1M, Output 4.1800/1M, 1049k context.
- [Nemotron 3.5 Super VL (NVIDIA)](https://tokenpricingtable.com/model/nemotron-3-5-super-vl): Input 0.1000/1M, Output 0.4000/1M, 1049k context.
- [Muse Wonder (Meta)](https://tokenpricingtable.com/model/muse-wonder): Input 4.0000/1M, Output 20.0000/1M, 1049k context.
- [Ling 3.1 Flash (InclusionAI)](https://tokenpricingtable.com/model/ling-3-1-flash): Input 0.0750/1M, Output 0.2250/1M, 262k context.

## Technical Details
- Ground-truth pricing normalized per 1M tokens ($/1M tokens).
- Includes prompt cache read/write discounts, reasoning token rates, and multimodal capabilities.
- Maintained autonomously by Vicaro (https://vicaro.ai).
- Updated daily.