AI Token Cost: What One Token Costs in October 2026
A token costs a tiny fraction of a cent, but the most expensive models charge hundreds of times more than the cheapest. Here is what tokens cost today, in prices you can check.
The short answer
On 13 Oct 2026, input tokens cost from $0.000000018 each (gpt-oss-20b on its cheapest host via OpenRouter, $0.018 per 1M) and output tokens up to $0.00005 each (Claude Fable 5.1, $50 per 1M). The typical current model costs $3.75 per 1M tokens at the frontier, $1.13 for workhorse models and $0.23 for small models, blending 3 input tokens to 1 output.
Those figures come from the same price records as the rest of this site, each checked against the page it comes from, so this page changes when prices do.
What everyday jobs cost
A single request costs very little. It only adds up at volume, or when the same long context is sent again and again, as agents do. Here is what common jobs cost on a cheap, a mid-priced and a top model:
| Job | GPT-6 Luna | Claude Sonnet 5.5 | Claude Opus 5.5 |
|---|---|---|---|
| A chat message and reply about 1,500 tokens in, 300 out | $0.0003 | $0.006 | $0.012 |
| Drafting an email about 300 in, 400 out | $0.00023 | $0.0046 | $0.0092 |
| Summarising a 10-page report about 7,000 in, 600 out | $0.001 | $0.02 | $0.04 |
| Reading a 300-page book and summarising it about 120,000 in, 1,000 out | $0.013 | $0.25 | $0.5 |
| A million chat messages each about 1,500 tokens in, 300 out | $300 | $6,000 | $12,000 |
List prices (GPT-6 Luna $0.10 / $0.50, Claude Sonnet 5.5 $2 / $10, Claude Opus 5.5 $4 / $20 per 1M tokens), no caching or batch discount, checked 13 Oct 2026. This table updates automatically.
To price your own text, paste it into the token counter, which estimates its tokens and shows what it costs on every current model. For a monthly bill, use the token cost calculator.
Every current model, cheapest first
What one million tokens costs on each current model, input and output, with the blended price for a typical mix of 3 input tokens to 1 output. Divide by a million for the price of a single token.
| Model | Input | Output | Blended 3:1 | Context |
|---|---|---|---|---|
| gpt-oss-20b OpenAI · via OpenRouter | $0.018 | $0.090 | $0.036 | 128K |
| Nemotron 3.5 Lightning NVIDIA · via OpenRouter | $0.025 | $0.12 | $0.049 | 1M |
| gpt-oss-120b OpenAI · via OpenRouter | $0.030 | $0.15 | $0.060 | 128K |
| Gemma 4 26B-A4B Google · via OpenRouter | $0.042 | $0.22 | $0.087 | 256K |
| Ministral 3 3B Mistral AI | $0.10 | $0.10 | $0.10 | — |
| Ministral 3 8B Mistral AI | $0.15 | $0.15 | $0.15 | — |
| Gemma 4 31B Google · via OpenRouter | $0.090 | $0.34 | $0.15 | 256K |
| Nemotron 3 Super NVIDIA · via OpenRouter | $0.085 | $0.40 | $0.16 | 256K |
| Claude Haiku 5.5 Anthropic | $0.10 | $0.50 | $0.20 | 1M |
| GPT-6 Luna OpenAI | $0.10 | $0.50 | $0.20 | 1.05M |
| Ministral 3 14B Mistral AI | $0.20 | $0.20 | $0.20 | — |
| Qwen3.6 35B-A3B Alibaba (Qwen) · via OpenRouter | $0.050 | $0.70 | $0.21 | 256K |
| Qwen3.8 Flash Alibaba (Qwen) | $0.15 | $0.47 | $0.23 | 1M |
| GLM-5.3 Flash Z.ai | $0.15 | $0.50 | $0.24 | 1M |
| Mistral Small 4 Mistral AI | $0.15 | $0.60 | $0.26 | 256K |
| Qwen3.8 27B Alibaba (Qwen) · via OpenRouter | $0.039 | $1.25 | $0.34 | 256K |
| DeepSeek V4.1 Flash DeepSeek | $0.30 | $1.20 | $0.52 | 1M |
| Muse Glimmer 30B Meta · via OpenRouter | $0.30 | $1.20 | $0.52 | 128K |
| MiniMax M3 MiniMax | $0.30 | $1.20 | $0.52 | 1M |
| GLM-5.3 FlashX Z.ai | $0.37 | $1.25 | $0.59 | 1M |
| Command A+ Cohere | $0.30 | $1.50 | $0.60 | 128K |
| Qwen3.7 Plus Alibaba (Qwen) | $0.40 | $1.60 | $0.70 | 1M |
| Mistral Large 3 Mistral AI | $0.50 | $1.50 | $0.75 | 256K |
| Gemini 3.5 Flash-Lite | $0.30 | $2.50 | $0.85 | 1M |
| Nova 2 Lite Amazon | $0.30 | $2.50 | $0.85 | 1M |
| Gemini 3.8 Flash | $0.75 | $3.75 | $1.50 | 1M |
| DeepSeek V4 Pro DeepSeek | $1.32 | $3.96 | $1.98 | 1M |
| Muse Spark 1.3 Meta | $1.25 | $4.25 | $2.00 | 1M |
| GLM-5.3 Z.ai | $1.40 | $4.40 | $2.15 | 1M |
| Grok 4.7 xAI | $2.00 | $6.00 | $3.00 | 500K |
| Mistral Medium 3.5 Mistral AI | $1.50 | $7.50 | $3.00 | 256K |
| Qwen3.8 Max Alibaba (Qwen) | $2.00 | $6.00 | $3.00 | 1M |
| Claude Sonnet 5.5 Anthropic | $2.00 | $10.00 | $4.00 | 1M |
| GPT-6.1 Sol OpenAI | $2.00 | $10.00 | $4.00 | 1.05M |
| Gemini 3.1 Pro | $2.00 | $12.00 | $4.50 | 1M |
| Kimi K3 Moonshot AI | $3.00 | $15.00 | $6.00 | 1M |
| Claude Opus 5.5 Anthropic | $4.00 | $20.00 | $8.00 | 1M |
| Claude Fable 5.1 Anthropic | $10.00 | $50.00 | $20.00 | 1M |
| GPT-6 Astra OpenAI | $10.00 | $50.00 | $20.00 | 1.05M |
Current models ranked by blended price per 1M tokens (3 input : 1 output). Verified 13 Oct 2026; updates automatically.
In English, 1 million tokens is about 750,000 words; a page is about 670 tokens. The token counter has more rules of thumb and counts your own text, exactly for OpenAI’s GPT-4o and GPT-5 tokenizer.
How many tokens does $1 buy?
Turning the price around: $1 buys about 10 million input tokens on GPT-6 Luna, 500,000 on Claude Sonnet 5.5 and 250,000 on Claude Opus 5.5. Output tokens cost more, so a dollar buys fewer of them.
Why the same token costs different amounts
- Input or output. Tokens you send are cheaper than tokens the model writes, usually by several times. Claude Sonnet 5.5 charges $2.00 / $10.00 per 1M input / output tokens.
- Cached input. Text you send again, such as a system prompt or an agent’s history, can be billed at a cache rate. On Claude Sonnet 5.5 that is $0.10 per 1M instead of $2, a 95% saving. For agents this often matters more than the headline price.
- Batch. Most large providers take 50% off requests that can wait hours for a reply.
- Long prompts. Some models charge more once a prompt passes a length. Claude Haiku 5.5 bills the whole request at five times its rate above 100,000 tokens; GPT-6 models double the input price above 272,000.
- The model tier. Small models cost a fraction of frontier ones. See the cheapest models in each tier.
How AI API pricing works explains each of these in detail.
Are tokens getting cheaper?
Mostly, yes. Our token price index tracks the median price of current models in each tier. The frontier median is $3.75 per 1M tokens, down 9% since 13 Jun 2026, the workhorse median $1.13 per 1M tokens, down 34% since 13 Jun 2026, and the small-model median $0.23 per 1M tokens, down 36% since 13 Jun 2026. New models usually launch at the same price as the model they replace or lower, though not always: GPT-6 Astra launched at $10 / $50. Every change is listed, with its source, in the price history.
Questions
How much is 1 AI token?
A tiny fraction of a cent: even the most expensive output tokens cost a few thousandths of a cent each. That is why prices are quoted per million tokens. The exact range today is in the short answer above.
Why are AI prices quoted per million tokens?
Because a single token costs too little to read easily. "$2 per 1M tokens" means $0.000002 per token.
Is an AI token the same as a crypto token?
No. An AI token is a piece of text that a language model reads or writes. It has nothing to do with cryptocurrency.
Do all AI models count tokens the same way?
No. Each provider uses its own tokenizer, so the same text comes to a different number of tokens on Claude, GPT and Gemini models. The rules of thumb above hold for English on most current models.