Home / Model Watch / GPT-6 Astra
GPT-6 Astra: price, access and what people say
GPT-6 Astra is OpenAI’s most capable model, out since September 2026. Through the API it costs $10.00 per million input tokens and $50.00 per million output tokens, five times as much as GPT-6.1 Sol. Here is what OpenAI says it is for, how to get it, what it costs in practice and how developers reacted.
Checked October 18, 2026 against OpenAI’s GPT-6 Astra model page
What GPT-6 Astra is for
OpenAI calls GPT-6 Astra its most capable model for the most demanding work: complex reasoning, coding, computer use, research and document creation. It reads text and images and writes text, has five reasoning settings from low to max, and knows the world up to April 30, 2026. To balance intelligence and cost, OpenAI points to GPT-6.1 Sol instead, and to GPT-6 Luna for cost-sensitive, high-volume work.
According to Wikipedia, OpenAI unveiled it as a limited preview on September 3, 2026 and opened it to paid ChatGPT users the next day, in a version that refuses some prompts in areas such as cybersecurity. GPT-6 Sol and GPT-6 Luna followed on September 22. OpenAI’s API changelog also dates Astra’s release for developers to September 3.
How much GPT-6 Astra costs
In ChatGPT, GPT-6 Astra comes with Plus ($20 a month) and Pro (from $100, with more use), but not with Free or Go; OpenAI’s Pro plan uses it for “Pro reasoning”. Pro has tiers of $100, $200 and $500 in Apple’s App Store. Through OpenAI’s API you pay per token:
| Standard | Prompts over 272,000 tokens | Batch and Flex | |
|---|---|---|---|
| Input | $10.00 | $20.00 | $5.00 |
| Cached input | $1.00 | $2.00 | $0.50 |
| Cache writes | $12.50 | $25.00 | — |
| Output | $50.00 | $75.00 | $25.00 |
Fast mode costs twice the applicable rate. Prompts over 272,000 input tokens are billed at the higher rate for the whole request.
What it costs in practice
Simon Willison asks every new model to draw a pelican riding a bicycle and posted the result and cost in the launch discussion. Higher reasoning settings use more output tokens, so they cost more, and they aren’t always better: one commenter noted that in OpenAI’s own scores, max effort did slightly worse than high on two coding benchmarks.
GPT-6 Astra for your use
Choose models
- Turn on JavaScript to see the costs. The price tables below work without it.
Cheaper models that score about as well as GPT-6 Astra
| Model | Company | Score | Points vs GPT-6 Astra | Blended per 1M | Costs less by |
|---|---|---|---|---|---|
| GPT-6.1 Sol | OpenAI | 166.3 | −0.1 | $4.00 | 80% |
| Gemini 4 Argon | 164.6 | −1.8 | $4.00 | 80% | |
| Claude Sonnet 5.5 | Anthropic | 164.4 | −2.0 | $4.00 | 80% |
| Claude Opus 5.5 | Anthropic | 166.8 | +0.4 | $8.00 | 60% |
Models within 3 points of GPT-6 Astra’s 166.4 on the Epoch Capabilities Index (CC BY 4.0), whose blended price (3 input tokens to 1 output) is lower than its $20.00. A few points is within the index’s margin of error; test on your own work before you switch.
Capability for the price
Epoch Capabilities Index (up) against the blended price per 1M tokens, 3 input to 1 output (right, log scale). Dots on the line have no model that is both cheaper and higher scoring. Hover or tap a dot for its numbers.
Show the numbers (16 models)
| # | Model | Company | Score | Blended per 1M | Value line |
|---|---|---|---|---|---|
| 1 | Claude Opus 5.5 | Anthropic | 166.8 | $8.00 | Yes |
| 2 | GPT-6 Astra | OpenAI | 166.4 | $20.00 | |
| 3 | GPT-6.1 Sol | OpenAI | 166.3 | $4.00 | Yes |
| 4 | Claude Fable 5.1 | Anthropic | 164.7 | $20.00 | |
| 5 | Gemini 4 Argon | 164.6 | $4.00 | ||
| 6 | Claude Sonnet 5.5 | Anthropic | 164.4 | $4.00 | |
| 7 | GPT-6 Sol | OpenAI | 162.6 | $4.00 | |
| 8 | Claude Opus 5 | Anthropic | 162.6 | $10.00 | |
| 9 | GPT-5.6 Sol | OpenAI | 161.7 | $8.00 | |
| 10 | GPT-5.6 Terra | OpenAI | 159.4 | $4.50 | |
| 11 | Claude Haiku 5.5 | Anthropic | 159.0 | $0.20 | Yes |
| 12 | Gemini 3.8 Flash | 157.2 | $1.50 | ||
| 13 | GPT-6 Luna | OpenAI | 156.3 | $0.20 | |
| 14 | DeepSeek Flash | DeepSeek | 155.6 | $0.525 | |
| 15 | DeepSeek V4 Pro | DeepSeek | 155.3 | $1.98 | |
| 16 | Grok 4.7 | xAI | 154.4 | $3.00 |
Scores: Epoch AI, Capabilities & benchmarking (Epoch Capabilities Index), CC BY 4.0, checked October 18, 2026. Each score carries a margin of error of about ±3 points, so models a few points apart may be level. Prices: each company’s pricing page, standard tier.
How to get GPT-6 Astra
- In ChatGPT: on a paid plan.
- Through OpenAI’s API: model ID
gpt-6-astra, with the Responses, Chat Completions and Batch endpoints. - On Amazon Bedrock: AWS made it generally available through its own APIs.
What people say
GPT-6 Astra’s launch discussion on Hacker News drew more than 2,000 comments, one of the biggest of the year.
- Benchmarks. OpenAI’s tables put Astra ahead of Anthropic’s Fable and Opus on most tests and near the top of the ARC-AGI-3 puzzle benchmark. Commenters questioned the custom setup behind that score, and pointed out that an independent index ranked Astra lower.
- Switching. Some developers said they were leaving Claude Max for OpenAI’s Codex because of usage limits; others said speed now matters more to them than raw intelligence.
- The “AGI” talk. Many were unconvinced by the promotional demos, which show ordinary tasks such as uploading photos and changing slide colors.
Questions
How much is GPT-6 Astra?
Through the API, $10.00 per million input tokens and $50.00 per million output tokens, half that with Batch or Flex. In ChatGPT it comes with Plus ($20 a month) and Pro, not with Free or Go.
Is GPT-6 Astra out?
Yes. It reached paid ChatGPT users on September 4, 2026, and today comes with the Plus and Pro plans. It is also on OpenAI’s API and Amazon Bedrock.
How good is GPT-6 Astra?
OpenAI calls it its most capable model and its own benchmark tables put it ahead of Anthropic’s top models on most tests; independent rankings are more mixed. Anthropic, for its part, says Claude Opus 5.5 matches it at about 40% of the cost. Test it on your own work before paying for the top tier.
GPT-6 Astra or GPT-6.1 Sol?
Sol costs a fifth as much and OpenAI describes it as near-Astra performance. See GPT-6.1 Sol vs GPT-6 Astra.
Sources
- OpenAI: GPT-6 Astra (purpose, limits, prices, endpoints)
- Wikipedia: GPT-6 Astra and OpenAI’s API changelog (release dates)
- AWS: GPT-6 Astra on Amazon Bedrock
- ChatGPT pricing (plan prices) and ChatGPT on the App Store (Pro tiers)
- Anthropic: Claude Opus 5.5 (its comparison with GPT-6 Astra)
- Hacker News: GPT-6 Astra launch discussion