AI price index
AI model pricing, compared.
Every model’s price per million tokens, the cache and batch discounts, and what a real workload costs a month. 340 models, refreshed daily.
Prices from OpenRouter’s models API, updated October 29, 2026
Pricing by provider
Claude API pricing
Claude Fable, Opus, Sonnet and Haiku: the models most coding agents run on.
15 models · from $0.10 per 1M inputOpenAI API pricing
GPT models, from the tiny ones to the pro reasoning tiers.
58 models · from $0.018 per 1M inputGemini API pricing
Gemini Flash, Pro and the image models, plus open Gemma.
28 models · from $0.05 per 1M inputDeepSeek API pricing
The price floor for strong open-weight models, with very cheap cache reads.
14 models · from $0.045 per 1M inputGrok API pricing
Grok models from xAI, including the fast variants.
7 models · from $1 per 1M inputMistral API pricing
European models from small to large, plus Codestral for code.
20 models · from $0.029 per 1M inputQwen API pricing
Qwen open and hosted models, from tiny to frontier-sized.
41 models · from $0.03 per 1M inputLlama API pricing
Meta’s Llama models, served by many hosts at different prices.
16 models · from $0.027 per 1M inputKimi API pricing
Kimi models from Moonshot AI, popular for long agentic runs.
8 models · from $0.239 per 1M inputGLM API pricing
GLM models from Z.ai, a cheap option for coding agents.
16 models · from $0.006 per 1M inputMiniMax API pricing
MiniMax text models with very long context.
9 models · from $0.15 per 1M inputThe newest models from each provider
| Model | Provider | Input /1M | Output /1M | Cache read /1M | Context | Chatbot, per month* |
|---|---|---|---|---|---|---|
| Claude Opus 5.5 | Anthropic | $4 | $20 | $0.20 | 1M | $420 |
| Muse Wonder | Meta | $4 | $20 | $0.20 | 1.05M | $420 |
| Qwen3.8 Max Prime | Alibaba Qwen | $4 | $12 | $0.50 | 1M | $324 |
| GLM 5.3 Prime | Z.ai | $2.80 | $8.80 | $0.56 | 1M | $232 |
| Claude Sonnet 5.5 | Anthropic | $2 | $10 | $0.10 | 1M | $210 |
| GPT-6.1 Sol | OpenAI | $2 | $10 | $0.10 | 1.05M | $210 |
| Gemini 4 Argon | $2 | $10 | $0.10 | 1.05M | $210 | |
| Kimi K3.1 Pro | Moonshot AI | $2 | $10 | $0.20 | 1.05M | $210 |
| Nano Banana 2.1 | $1.50 | $7.50 | – | 66K | $158 | |
| Grok 4.7 | xAI | $2 | $6 | $0.50 | 500K | $162 |
| Grok 4.6 | xAI | $2 | $6 | $0.50 | 500K | $162 |
| Grok 4.5 | xAI | $2 | $6 | $0.30 | 500K | $162 |
| Mistral Medium 3.5 | Mistral | $1.50 | $7.50 | – | 262K | $158 |
| Qwen3.8 Max (0902) | Alibaba Qwen | $2 | $6 | $0.25 | 1M | $162 |
| Kimi K3 | Moonshot AI | $0.239 | $10.80 | $0.21 | 1.05M | $140 |
| GPT-6.1 Terra | OpenAI | $1 | $5 | $0.05 | 1.05M | $105 |
| GPT-6.1 Terra Pro | OpenAI | $1 | $5 | $0.05 | 1.05M | $105 |
| Muse Spark Open | Meta | $1.25 | $4.25 | $0.15 | 1.05M | $107 |
| Muse Spark 1.3 | Meta | $1.25 | $4.25 | $0.15 | 1.05M | $107 |
| Gemini 3.8 Flash | $0.825 | $4.13 | $0.083 | 1.05M | $86.62 | |
| Kimi K2.7 Code | Moonshot AI | $0.56 | $3.10 | $0.15 | 262K | $62.40 |
| Mistral Large 4 | Mistral | $0.70 | $2.10 | $0.07 | 1.05M | $56.70 |
| GLM 5.3 FlashX | Z.ai | $0.37 | $1.25 | $0.09 | 1.05M | $31.65 |
| MiniMax M3 | MiniMax | $0.30 | $1.20 | $0.06 | 1.05M | $27.90 |
| MiniMax M2.7 | MiniMax | $0.30 | $1.20 | $0.06 | 205K | $27.90 |
| DeepSeek V4 Flash Vision Exp | DeepSeek | $0.216 | $0.647 | $0.0069 | 1.05M | $17.46 |
| DeepSeek V4.1 Flash 1017 | DeepSeek | $0.15 | $0.60 | $0.003 | 1.05M | $13.95 |
| Mistral Small 4 | Mistral | $0.15 | $0.60 | $0.015 | 262K | $13.95 |
| MiniMax M3.1 Flash | MiniMax | $0.15 | $0.60 | $0.015 | 1.05M | $13.95 |
| Qwen3.8 Omni Flash | Alibaba Qwen | $0.15 | $0.47 | $0.016 | 1M | $12.39 |
| Claude Haiku 5.5 | Anthropic | $0.10 | $0.50 | $0.01 | 1M | $10.50 |
| DeepSeek V4.1 Flash | DeepSeek | $0.045 | $0.50 | $0.006 | 1.05M | $8.03 |
| GLM 5.3 Flash | Z.ai | $0.048 | $0.10 | $0.044 | 1.05M | $3.36 |
* Support chatbot: 1,000 chats a day, 1,500 tokens in, 400 out, 30 days, no caching. Prices in US dollars per 1M tokens.
Work out your own bill
About these prices
How are these prices collected?
From OpenRouter’s public models API, which lists the per-token price of every model it routes to. The site rebuilds from it every day; this page was built on October 29, 2026.
Are OpenRouter prices the same as buying from the provider directly?
For most models OpenRouter passes the provider’s list price through per token. Credit purchases carry a small platform fee, and some providers price long prompts or regions differently, so check the provider’s page for edge cases.
What does “per 1M tokens” mean?
A token is roughly ¾ of an English word. A price of $3 per 1M input tokens means a 1,000-token prompt costs $0.003. Output tokens (what the model writes) are usually 3–8× more expensive than input.
How do I see what I actually spend per model?
Connect an OpenRouter management key to Glow and you get your real spend by model, API key and product, with cache hit rate and cheaper-model suggestions. How it works.
See what you actually spend.
List prices are half the story. Connect OpenRouter and Glow shows your real spend by model, key and product, and which swap saves the most.
- Read-only keys
- Never shown again
- 2FA + Face ID
- 7-day trial