Tokens to cost
One million input tokens costs between $0.10 and $10 depending on the model — a 100× spread for the same volume. The table prices 1M tokens on every model in the dataset.
All models, 1M tokens
| # | Model | Provider | Input | Output | Blended |
|---|---|---|---|---|---|
| 1 | GPT-6 Luna | OpenAI | $0.10 | $0.50 | $0.225 |
| 2 | GPT-5.6 Luna | OpenAI | $0.20 | $1.20 | $0.50 |
| 3 | GPT-5.4 nano | OpenAI | $0.20 | $1.25 | $0.512 |
| 4 | Gemini 3.5 Flash-Lite | Google Gemini | $0.30 | $2.50 | $0.925 |
| 5 | DeepSeek V4.1 Flash | DeepSeek | $0.30 | $1.20 | $0.60 |
| 6 | GPT-5.4 mini | OpenAI | $0.75 | $4.50 | $1.88 |
| 7 | Gemini 3.8 Flash | Google Gemini | $0.75 | $3.75 | $1.69 |
| 8 | Gemini 3.7 Flash | Google Gemini | $0.75 | $3.75 | $1.69 |
| 9 | Claude Haiku 4.5 | Anthropic | $1 | $5 | $2.25 |
| 10 | Gemini 2.5 Pro | Google Gemini | $1.25 | $10 | $3.75 |
| 11 | Grok 4.3 | xAI Grok | $1.25 | $2.50 | $1.88 |
| 12 | DeepSeek V4 Pro | DeepSeek | $1.32 | $3.96 | $2.31 |
| 13 | Gemini 3.5 Flash | Google Gemini | $1.50 | $9 | $3.75 |
| 14 | GPT-5.2 | OpenAI | $1.75 | $14 | $5.25 |
| 15 | GPT-6.1 Sol | OpenAI | $2 | $10 | $4.50 |
| 16 | GPT-6 Sol | OpenAI | $2 | $10 | $4.50 |
| 17 | GPT-5.6 Terra | OpenAI | $2 | $12 | $5 |
| 18 | GPT-4.1 | OpenAI | $2 | $8 | $4 |
| 19 | Claude Sonnet 5 | Anthropic | $2 | $10 | $4.50 |
| 20 | Gemini 3.1 Pro (Preview) | Google Gemini | $2 | $12 | $5 |
| 21 | Grok 4.6 | xAI Grok | $2 | $6 | $3.50 |
| 22 | GPT-5.4 | OpenAI | $2.50 | $15 | $6.25 |
| 23 | Claude Sonnet 4.6 | Anthropic | $3 | $15 | $6.75 |
| 24 | GPT-5.6 Sol | OpenAI | $4 | $20 | $9 |
| 25 | GPT-5.5 | OpenAI | $5 | $30 | $12.50 |
| 26 | Claude Opus 5 | Anthropic | $5 | $25 | $11.25 |
| 27 | GPT-6 Astra | OpenAI | $10 | $50 | $22.50 |
| 28 | Claude Fable 5.1 | Anthropic | $10 | $50 | $22.50 |
Input = 1,000,000 prompt tokens. Output = 1,000,000 generated tokens. Blended = input + 25% of output, the same weighting used across this site.
Scale it linearly
| Volume | GPT-6 Luna | Median model | GPT-6 Astra |
|---|---|---|---|
| 10,000 | $0.001 | $0.02 | $0.10 |
| 100,000 | $0.01 | $0.20 | $1 |
| 1,000,000 | $0.10 | $2 | $10 |
| 10,000,000 | $1 | $20 | $100 |
| 100,000,000 | $10 | $200 | $1,000 |
Input tokens only, at list price. Prompt caching, batch discounts and long-context surcharges all move these numbers — the model pages show each effect separately.
FAQ
How much does 1 million tokens cost?
From $0.10 on GPT-6 Luna to $10 on GPT-6 Astra for input, and from $0.50 to $50 for output. The spread is 100×, which is why model choice dominates every other cost decision.
How much is 1,000 tokens?
At the cheapest rate in the dataset, $0.0001. At the most expensive, $0.01. Under a cent either way is the practical answer.
Does input or output cost more?
Output, on every model in this dataset — between 3× and 5× the input rate. A chat that reads 4,000 tokens and writes 600 is still dominated by the prompt, but a long generation is dominated by the completion.
Keep going
- Every model ranked by price
- Head-to-head cost comparisons
- What 1,000 words costs
- Recent API price changes
- All 28 models ranked by price
- Run your own volumes in the calculator