GPT-6 Pricing: Astra $10 / $50, Sol $2 / $10, Luna $0.10 / $0.50
OpenAI's GPT-6 family arrived in two steps in September. The top model is OpenAI's most expensive standard model yet; the two smaller ones cost less than the GPT-5.6 models they replace.
OpenAI launched GPT-6 Astra on 3 September 2026, with broad availability from 4 September, and followed with GPT-6 Sol and GPT-6 Luna on 22 September. All three have a 1.05M-token context window, cached input at a tenth of the input price, and 50% off through the Batch API.
GPT-6 prices
| Model | Input | Cached input | Output |
|---|---|---|---|
| GPT-6 Astra | $10.00 | $1.00 | $50.00 |
| GPT-6 Sol | $2.00 | $0.20 | $10.00 |
| GPT-6 Luna | $0.10 | $0.01 | $0.50 |
Per 1M tokens. Prompts over 272K tokens cost twice the input price and 1.5 times the output price for the whole request, the same rule as GPT-5.6. Live prices are on our GPT-6 pricing page.
Compared with GPT-5.6
| GPT-5.6 model | Price | GPT-6 model | Price |
|---|---|---|---|
| GPT-5.6 Sol | $4.00 / $20.00 | GPT-6 Astra | $10.00 / $50.00 |
| GPT-5.6 Terra | $2.00 / $12.00 | GPT-6 Sol | $2.00 / $10.00 |
| GPT-5.6 Luna | $0.20 / $1.20 | GPT-6 Luna | $0.10 / $0.50 |
GPT-6 Sol matches GPT-5.6 Terra on input and costs 17% less on output. GPT-6 Luna costs half as much as GPT-5.6 Luna on input and 58% less on output, and GPT-5.6 Luna had already been cut by 80% in July. At the top, GPT-6 Astra replaces GPT-5.6 Sol and costs two and a half times its promotional $4 / $20.
At $10 / $50, GPT-6 Astra is OpenAI's most expensive standard model and costs the same as Claude Fable 5.1. The GPT-5.6 models remain available, and GPT-5.6 Sol's promotional price is guaranteed until at least 21 November 2026.
What it costs in practice
| Model | Input / output per 1M | Per month |
|---|---|---|
| GPT-6 Luna | $0.10 / $0.50 | $2.00 |
| Gemini 3.8 Flash | $0.75 / $3.75 | $15.00 |
| GPT-6 Sol | $2.00 / $10.00 | $40.00 |
| Claude Sonnet 5 | $2.00 / $10.00 | $40.00 |
| Claude Opus 5.5 | $4.00 / $20.00 | $80.00 |
| GPT-6 Astra | $10.00 / $50.00 | $200 |
10M input and 2M output tokens a month. Prices verified 27 Sep 2026; this table updates automatically.
For cache-heavy agents, the tenth-of-input cache rate matters more than the headline price. A coding agent with 750M input tokens a month, 85% of them cached, and 8M output tokens costs $433 on GPT-6 Sol and $2,163 on GPT-6 Astra.
Free daily tokens
OpenAI's data-sharing programme includes the new models. Eligible organisations that agree to share API inputs and outputs with OpenAI get up to 1M free tokens a day across GPT-6 Astra, Sol and Luna and GPT-5.6 Sol (250K a day on usage tiers 1 and 2). Read OpenAI's terms before opting in; it is not suitable for confidential data.
Which GPT-6 model to use
- GPT-6 Luna for high-volume, simple work: classification, extraction, routing, short answers. It is one of the cheapest models from a major lab.
- GPT-6 Sol for most production work, including chat, agents and coding. Its price matches Claude Sonnet 5.
- GPT-6 Astra for the hardest tasks, where a better first attempt saves more than the extra cost. On the Arena text leaderboard on 25 September it scored 1478, below Claude Opus 5.5 (1509, provisional) and Gemini 3.8 Flash (1492), so test it on your own tasks before paying five times Sol's price.
Compare them with rivals: Claude Opus 5.5 vs GPT-6 Astra, Claude Sonnet 5 vs GPT-6 Sol, GPT-6 Luna vs Claude Haiku 4.5.
Sources: OpenAI API pricing, GPT-6 Astra model page.