OpenAI API pricing 2026
OpenAI API pricing runs from $0.10 per million input tokens for GPT-6 Luna to $30 for GPT-5.5 Pro, with output tokens costing 5 to 6 times as much as input. Cached input is billed at a tenth of the input price or less on most models. Prices below are the standard tier for short prompts, from OpenAI's pricing page.
OpenAI prices per 1M tokens
| Model | Input | Cached | Output | Context |
|---|---|---|---|---|
GPT-6 Astragpt-6-astraOver 272K tokens: $20.00 in, $75.00 out. | $10.00 | $1.00 | $50.00 | 1.05M |
GPT-6.1 Solgpt-6.1-solOver 272K tokens: $4.00 in, $15.00 out. | $2.00 | $0.10 | $10.00 | 1.05M |
GPT-6.1 Terragpt-6.1-terraOver 272K tokens: $2.00 in, $7.50 out. | $1.00 | $0.05 | $5.00 | 1.05M |
GPT-6 Solgpt-6-solOver 272K tokens: $4.00 in, $15.00 out. | $2.00 | $0.20 | $10.00 | Not listed |
GPT-6 Lunagpt-6-luna | $0.10 | $0.01 | $0.50 | 1.05M |
GPT-5.6 Solgpt-5.6-solOpenAI says its promotional pricing runs at least through 21 Nov 2026. | $4.00 | $0.40 | $20.00 | Not listed |
GPT-5.6 Terragpt-5.6-terraOver 272K tokens: $4.00 in, $18.00 out. Max input 922K tokens. | $2.00 | $0.20 | $12.00 | 1.05M |
GPT-5.6 Lunagpt-5.6-lunaOver 272K tokens: $0.40 in, $1.80 out. | $0.20 | $0.02 | $1.20 | Not listed |
GPT-5.5gpt-5.5Over 272K tokens: $10.00 in, $45.00 out. | $5.00 | $0.50 | $30.00 | Not listed |
GPT-5.5 Progpt-5.5-pro | $30.00 | Not listed | $180.00 | 1.05M |
GPT-5.4gpt-5.4Over 272K tokens: $5.00 in, $22.50 out. | $2.50 | $0.25 | $15.00 | 1.05M |
GPT-5.4 minigpt-5.4-miniMax input 272K tokens. | $0.75 | $0.075 | $4.50 | 400K |
GPT-5.4 nanogpt-5.4-nanoMax input 272K tokens. | $0.20 | $0.02 | $1.25 | 400K |
Standard tier, US dollars. Source: OpenAI API pricing, checked November 1, 2026.
OpenAI cost calculator
Includes reasoning or thinking tokens.
Cache write fees are not included.
Estimate tokens from text
About 0 tokens
A rough count at about 4 characters per token for English. Every provider tokenizes differently; Anthropic says Claude 4.7 and later produce about 30% more tokens for the same text.
| Model | Per request | Per month |
|---|---|---|
| GPT-6 Luna$0.10 in / $0.50 out | $0.0011 | $10.50 |
| GPT-5.6 Luna$0.20 in / $1.20 out | $0.0022 | $22.00 |
| GPT-5.4 nano$0.20 in / $1.25 out | $0.0022 | $22.25 |
| GPT-5.4 mini$0.75 in / $4.50 out | $0.0083 | $82.50 |
| GPT-6.1 Terra$1.00 in / $5.00 out | $0.011 | $105.00 |
| GPT-6.1 Sol$2.00 in / $10.00 out | $0.021 | $210.00 |
| GPT-6 Sol$2.00 in / $10.00 out | $0.021 | $210.00 |
| GPT-5.6 Terra$2.00 in / $12.00 out | $0.022 | $220.00 |
| GPT-5.4$2.50 in / $15.00 out | $0.028 | $275.00 |
| GPT-5.6 Sol$4.00 in / $20.00 out | $0.042 | $420.00 |
| GPT-5.5$5.00 in / $30.00 out | $0.055 | $550.00 |
| GPT-6 Astra$10.00 in / $50.00 out | $0.11 | $1,050.00 |
| GPT-5.5 Pro$30.00 in / $180.00 out | $0.33 | $3,300.00 |
Standard tier prices per 1M tokens. Batch processing is about half price at OpenAI, Anthropic and Google. Taxes, tool fees and free tiers are not included.
Caching, batch and other pricing rules
- Prompts over 272K input tokens cost more on long context models. On every model where we show a long prompt price, the page lists 2x input and 1.5x output.
- The Batch table lists half the standard price on the rows we compared.
- Flex and Fast tiers also exist. OpenAI renamed Priority processing to Fast mode on 30 July 2026; it costs more than standard.
- Regional and data residency endpoints add 10%.
- The Pro models have no cached input discount.
Price change log
- Rechecked: no price changes. Added GPT-6.1 Terra, GPT-6 Sol and GPT-5.6 Luna, and long prompt prices for GPT-6 Astra, GPT-6.1 Sol and GPT-5.5.
- First check. GPT-5.6 Sol is on promotional pricing until at least 21 Nov 2026.
FAQ
What is the cheapest OpenAI API model?
GPT-6 Luna, at $0.10 per million input tokens and $0.50 per million output tokens on the standard tier, with a 1.05M token context window.
How much does GPT-6 Astra cost?
GPT-6 Astra costs $10 per million input tokens, $1 per million cached input tokens and $50 per million output tokens on the standard tier.
How do I lower my OpenAI API bill?
Use a smaller model where quality allows, put stable instructions at the start of the prompt so they hit the cache, send work that can wait through the Batch API at about half price, and cap output length, since output tokens cost several times more than input.