Key takeaways
- Claude Haiku 5.5 is the cheapest newest model in the table at $0.10 input and $0.50 output per million tokens; Claude Fable 5.1 costs $10 and $50.
- Anthropic and Google prices were checked on their pricing pages on November 6, 2026; OpenAI's page was unreadable, so its prices are OpenRouter list prices.
- OpenRouter lists Gemini 3.8 Flash at $0.825 / $4.13, while Google lists $0.75 / $3.75.
AI API pricing in one place, as of November 6, 2026: the cheapest model in the newest-model table below is Anthropic's Claude Haiku 5.5 at $0.10 per million input tokens and $0.50 per million output tokens. The most expensive Anthropic model is Claude Fable 5.1 at $10 input and $50 output, according to Anthropic's pricing page. All prices are in US dollars per million tokens.
OpenAI's pricing page refused our request (HTTP 403), so the OpenAI prices here are OpenRouter list prices, not OpenAI's own. Anthropic and Google prices were checked on their pricing pages. Everything else comes from OpenRouter's model list.
The newest model from each company
| Company | Model | Input | Output | Listed on OpenRouter |
|---|---|---|---|---|
| OpenAI | GPT-6.1 Terra Pro | $1.00 | $5.00 | 2026-10-26 |
| Anthropic | Claude Haiku 5.5 | $0.10 | $0.50 | 2026-10-07 |
| Gemini 4 Argon | $2.00 | $10.00 | 2026-10-13 | |
| xAI | Grok 4.7 | $2.00 | $6.00 | 2026-09-21 |
| DeepSeek | DeepSeek V4.1 Flash 1017 | $0.15 | $0.60 | 2026-10-17 |
| Mistral | Mistral Large 4 | $0.70 | $2.10 | 2026-10-06 |
| Meta | Muse Spark Open | $1.25 | $4.25 | 2026-10-15 |
| Alibaba (Qwen) | Qwen3.8 Max Prime | $4.00 | $12.00 | 2026-09-23 |
| Moonshot AI | Kimi K3.1 Pro | $2.00 | $10.00 | 2026-10-23 |
| Z.ai | GLM 5.3 FlashX Legal | $0.37 | $1.25 | 2026-11-02 |
| MiniMax | MiniMax M3.1 Flash | $0.15 | $0.60 | 2026-10-19 |
| Amazon | Nova 2 Lite | $0.30 | $2.50 | 2025-12-02 |
| Cohere | Command A+ | $0.30 | $1.50 | 2026-09-22 |
| NVIDIA | Nemotron 3.5 Super VL | $0.15 | $0.50 | 2026-10-15 |
"Newest" means the most recently listed on OpenRouter, which is not always the company's best model. Anthropic's newest listing is its smallest model, and its flagship is Fable 5.1.
Claude API pricing
Checked on Anthropic's pricing page on November 6, 2026. Anthropic lists the 5-minute cache write price too: $12.50 for Fable 5.1, $5 for Opus 5.5, $2.50 for Sonnet 5.5 and $0.125 for Haiku 5.5. Batch prices are half the standard price. Haiku 5.5 has a higher tier for prompts over 100K tokens: $0.50 input, $2.50 output and $0.05 cache read. Anthropic's page lists up to 1M tokens of context across its models.
| Model | Input | Output | Cached input (read) | Batch (input / output) |
|---|---|---|---|---|
| Claude Fable 5.1 | $10.00 | $50.00 | $0.25 | $5.00 / $25.00 |
| Claude Opus 5.5 | $4.00 | $20.00 | $0.20 | $2.00 / $10.00 |
| Claude Sonnet 5.5 | $2.00 | $10.00 | $0.10 | $1.00 / $5.00 |
| Claude Haiku 5.5 (up to 100K tokens) | $0.10 | $0.50 | $0.01 | $0.05 / $0.25 |
Older models still on OpenRouter, from OpenRouter's list: Claude Opus 5 ($5.00 input, $25.00 output), Claude Sonnet 5 ($2.00 / $10.00), Claude Fable 5 ($10.00 / $50.00) and Claude Opus 4.8 ($5.00 / $25.00). Anthropic's page no longer lists them.
OpenAI API pricing
OpenAI's pricing page could not be read, so these are OpenRouter list prices. They may differ from what OpenAI charges directly.
| Model | Input | Output | Cached input | Batch (input / output) | Context |
|---|---|---|---|---|---|
| GPT-6.1 Terra Pro | $1.00 | $5.00 | $0.05 | $0.50 / $2.50 | 1.05M |
| GPT-6.1 Terra | $1.00 | $5.00 | $0.05 | $0.50 / $2.50 | 1.05M |
| GPT-6.1 Sol Pro | $2.00 | $10.00 | $0.10 | $1.00 / $5.00 | 1.05M |
| GPT-6.1 Sol | $2.00 | $10.00 | $0.10 | $1.00 / $5.00 | 1.05M |
| GPT-6 Luna Pro | $0.10 | $0.50 | $0.01 | $0.05 / $0.25 | 1.05M |
| GPT-6 Luna | $0.10 | $0.50 | $0.01 | $0.05 / $0.25 | 1.05M |
| GPT-6 Sol Pro | $2.00 | $10.00 | $0.20 | $1.00 / $5.00 | 1.05M |
| GPT-6 Sol | $2.00 | $10.00 | $0.20 | $1.00 / $5.00 | 1.05M |
Gemini API pricing
Checked on Google's Gemini API pricing page on November 6, 2026 (paid tier). Google notes that prices for Gemini 4 Argon, Gemini 3.8 Flash and Gemini 3.5 Flash-Lite rise on a later date, December 1, 2026 or January 1, 2027, so check the page before you commit to a budget.
| Model | Input | Output | Cached input | Batch (input / output) |
|---|---|---|---|---|
| Gemini 4 Argon | $2.00 | $10.00 | $0.10 | $1.00 / $5.00 |
| Gemini 3.8 Flash | $0.75 | $3.75 | $0.075 | $0.375 / $1.875 |
| Gemini 3.6 Flash | $0.75 | $3.75 | $0.075 | $0.375 / $1.875 |
| Gemini 3.5 Flash-Lite | $0.30 | $2.50 | $0.03 | $0.15 / $1.25 |
| Gemini 3.1 Pro Preview (prompts up to 200K) | $2.00 | $12.00 | $0.20 | $1.00 / $6.00 |
| Gemini 3.1 Pro Preview (prompts over 200K) | $4.00 | $18.00 | $0.40 | $2.00 / $9.00 |
| Gemini 3.1 Flash-Lite | $0.25 | $1.50 | $0.025 | $0.125 / $0.75 |
OpenRouter and Google disagree on one model. OpenRouter lists Gemini 3.8 Flash at $0.825 input and $4.13 output, while Google lists $0.75 and $3.75. The table uses Google's price. OpenRouter also lists Gemini 3.7 Flash ($0.75 / $3.75), Gemini 3.5 Flash ($1.50 / $9.00) and the open Gemma 4 26B A4B ($0.09 / $0.30), which we did not see on Google's page.
Grok API pricing
xAI's model docs are the company source. These are OpenRouter list prices.
| Model | Input | Output | Cached input | Batch (input / output) | Context |
|---|---|---|---|---|---|
| Grok 4.7 | $2.00 | $6.00 | $0.50 | $1.60 / $4.80 | 500K |
| Grok 4.6 | $2.00 | $6.00 | $0.50 | $1.60 / $4.80 | 500K |
| Grok 4.5 | $2.00 | $6.00 | $0.30 | 500K | |
| Grok Build 0.1 | $1.00 | $2.00 | $0.20 | 256K | |
| Grok 4.3 | $1.25 | $2.50 | $0.20 | $1.00 / $2.00 | 1M |
| Grok 4.20 Multi-Agent | $1.25 | $2.50 | $0.20 | 2M | |
| Grok 4.20 | $1.25 | $2.50 | $0.20 | 2M |
DeepSeek API pricing
DeepSeek's pricing page is the company source. These are OpenRouter list prices, and they vary a lot between versions of the same model because different providers serve them.
| Model | Input | Output | Cached input | Batch (input / output) | Context |
|---|---|---|---|---|---|
| DeepSeek V4.1 Flash 1017 | $0.15 | $0.60 | $0.003 | 1.05M | |
| DeepSeek V4.1 Flash | $0.051 | $0.567 | $0.018 | $0.112 / $0.336 | 1.05M |
| DeepSeek V4 Flash Vision Exp | $0.44 | $1.32 | $0.014 | 1.05M | |
| DeepSeek V4 Pro 0813 | $0.117 | $4.66 | $0.111 | 1.05M | |
| DeepSeek V4 Flash 0731 | $0.059 | $0.30 | $0.035 | 1.05M | |
| DeepSeek V4 Pro 0423 | $0.955 | $1.91 | $0.0796 | 1.05M | |
| DeepSeek V4 Flash 0423 | $0.091 | $0.182 | $0.0182 | 1.05M | |
| DeepSeek V3.2 | $0.26 | $0.38 | $0.13 | 164K |
Mistral API pricing
Mistral's pricing page is the company source. These are OpenRouter list prices.
| Model | Input | Output | Cached input | Batch (input / output) | Context |
|---|---|---|---|---|---|
| Mistral Large 4 | $0.70 | $2.10 | $0.07 | 1.05M | |
| Mistral Medium 3.5 | $1.50 | $7.50 | $0.75 / $3.75 | 262K | |
| Mistral Small 4 | $0.15 | $0.60 | $0.015 | 262K | |
| Devstral 2 2512 | $0.40 | $2.00 | $0.04 | 262K | |
| Ministral 3 14B 2512 | $0.20 | $0.20 | $0.02 | 262K | |
| Ministral 3 8B 2512 | $0.15 | $0.15 | $0.015 | 262K | |
| Ministral 3 3B 2512 | $0.10 | $0.10 | $0.01 | 131K | |
| Mistral Large 3 2512 | $0.50 | $1.50 | $0.05 | 262K |
Meta Muse API pricing
OpenRouter list prices.
| Model | Input | Output | Cached input | Context |
|---|---|---|---|---|
| Muse Spark Open | $1.25 | $4.25 | $0.15 | 1.05M |
| Muse Wonder | $4.00 | $20.00 | $0.20 | 1.05M |
| Muse Spark 1.3 Contributor | $0.10 | $0.20 | $0.002 | 1.05M |
| Muse Spark 1.3 | $1.25 | $4.25 | $0.15 | 1.05M |
| Muse Spark 1.2 Contributor | $0.10 | $0.20 | $0.002 | 1.05M |
| Muse Glimmer 30B | $0.30 | $1.20 | $0.04 | 131K |
| Muse Spark 1.2 | $1.25 | $4.25 | $0.15 | 1.05M |
| Muse Spark 1.1 | $1.25 | $4.25 | $0.15 | 1.05M |
Qwen API pricing
OpenRouter list prices.
| Model | Input | Output | Cached input | Batch (input / output) | Context |
|---|---|---|---|---|---|
| Qwen3.8 Max Prime | $4.00 | $12.00 | $0.50 | 1M | |
| Qwen3.8 Omni Flash | $0.15 | $0.47 | $0.016 | 1M | |
| Qwen3.8 Max (0902) | $2.00 | $6.00 | $0.25 | 1M | |
| Qwen3.8 Flash | $0.15 | $0.47 | $0.016 | 1M | |
| Qwen3.8 27B | $0.15 | $1.88 | $0.0375 | 1.02M | |
| Qwen3.8 2.4T A95B | $2.00 | $6.00 | $0.25 | 1.05M | |
| Qwen3.7 Flash | $0.03 | $0.13 | $0.006 | 1M | |
| Qwen3.7 Plus | $0.32 | $1.28 | $0.064 | $0.23 / $0.92 | 1M |
Kimi API pricing
Moonshot's pricing page is the company source. These are OpenRouter list prices.
| Model | Input | Output | Cached input | Batch (input / output) | Context |
|---|---|---|---|---|---|
| Kimi K3.1 Pro | $2.00 | $10.00 | $0.20 | 1.05M | |
| Kimi K3 | $0.192 | $16.00 | $0.182 | $2.28 / $11.40 | 1.05M |
| Kimi K2.7 Code | $0.60 | $3.50 | $0.16 | 262K | |
| Kimi K2.6 | $0.65 | $3.41 | $0.15 | 262K | |
| Kimi K2.5 | $0.50 | $2.50 | $0.15 | 262K | |
| Kimi K2 Thinking | $0.60 | $2.50 | $0.15 | 262K |
GLM API pricing
Z.ai's pricing page is the company source. These are OpenRouter list prices.
| Model | Input | Output | Cached input | Batch (input / output) | Context |
|---|---|---|---|---|---|
| GLM 5.3 FlashX Legal | $0.37 | $1.25 | $0.075 | 1.05M | |
| GLM 5.3 Prime | $2.80 | $8.80 | $0.56 | 1M | |
| GLM 5.3 FlashX | $0.37 | $1.25 | $0.075 | 1.05M | |
| GLM 5.3 Flash | $0.15 | $0.50 | $0.03 | $0.05 / $0.18 | 1.05M |
| GLM 5.3 | $0.0699 | $1.40 | $0.0689 | $0.45 / $2.00 | 1.05M |
| GLM 5.2 | $0.006 | $6.80 | $0.005 | 1.05M | |
| GLM 5.1 | $1.40 | $4.40 | $0.26 | 205K | |
| GLM 5V Turbo | $1.20 | $4.00 | $0.24 | 203K |
MiniMax, Amazon, Cohere and NVIDIA API pricing
OpenRouter list prices. Company pages: Amazon Bedrock and Cohere.
| Model | Input | Output | Cached input | Context |
|---|---|---|---|---|
| MiniMax M3.1 Flash | $0.15 | $0.60 | $0.015 | 1.05M |
| MiniMax M3 | $0.30 | $1.20 | $0.06 | 1.05M |
| MiniMax M2.7 | $0.21 | $0.84 | $0.042 | 205K |
| MiniMax M2.5 | $0.27 | $0.95 | $0.03 | 205K |
| Amazon Nova 2 Lite | $0.30 | $2.50 | 1M | |
| Cohere Command A+ | $0.30 | $1.50 | $0.15 | 192K |
| NVIDIA Nemotron 3.5 Super VL | $0.15 | $0.50 | $0.02 | 1.05M |
| NVIDIA Nemotron 3.5 Lightning | $0.07 | $0.20 | $0.04 | 262K |
| NVIDIA Nemotron 3 Ultra | $0.60 | $2.40 | $0.12 | 262K |
| NVIDIA Nemotron 3 Super | $0.075 | $0.40 | 1.05M | |
| NVIDIA Nemotron 3 Nano 30B A3B | $0.06 | $0.24 | $0.012 | 262K |
How to read these prices
- Input and output tokens. You pay separately for the text you send (input) and the text the model writes back (output). Output costs three to six times more per token. A token is a few characters, roughly three quarters of a word in English.
- Cached input. When many requests start with the same long prompt, the provider can store it and charge a lower price to read it again. Cached input is often a tenth of the normal input price. Some providers charge extra to write the cache the first time.
- Batch. If you can wait for results, batch APIs usually charge half the standard price for input and output.
- Long prompts. Some models charge more above a size limit, such as Gemini 3.1 Pro Preview above 200K tokens or Claude Haiku 5.5 above 100K.
- OpenRouter prices. OpenRouter passes through provider prices, so they can differ from the company's own page, especially for open models that several hosts serve.
Price changes
- November 6, 2026: first edition of this page. Anthropic and Google prices were checked against their pricing pages. OpenRouter lists Gemini 3.8 Flash at $0.825 input and $4.13 output, against Google's $0.75 and $3.75.
Frequently asked questions
How much does the Claude API cost?
As of November 6, 2026, Anthropic lists Claude Fable 5.1 at $10 input and $50 output per million tokens, Opus 5.5 at $4 and $20, Sonnet 5.5 at $2 and $10, and Haiku 5.5 at $0.10 and $0.50.
How much does the OpenAI API cost?
OpenAI's pricing page could not be read, so we use OpenRouter list prices: GPT-6.1 Sol is $2 input and $10 output per million tokens, GPT-6.1 Terra is $1 and $5, and GPT-6 Luna is $0.10 and $0.50.
How much does the Gemini API cost?
Google lists Gemini 4 Argon at $2 input and $10 output per million tokens on the paid tier, and Gemini 3.8 Flash at $0.75 and $3.75. Google says some prices rise on December 1, 2026 or January 1, 2027.
What is a token?
A token is a small chunk of text, a few characters long. English text averages roughly three quarters of a word per token. APIs bill by the million tokens.
Sources
- Claude pricingAnthropic
- Gemini API pricingGoogle
- OpenRouter models APIOpenRouter
- OpenAI API pricingOpenAI
- xAI models and pricingxAI
- DeepSeek models and pricingDeepSeek
- Mistral pricingMistral AI