Frontier flagship models
GPT-6 Astra
Made by OpenAI
OpenAI's flagship model for complex reasoning, coding and computer use — paid plan and API only.
What it is
GPT-6 Astra is the top of OpenAI's GPT-6 family, aimed at demanding reasoning, coding, computer-use and research work. It is reached through ChatGPT paid plans (Plus, Pro, Business, Enterprise) and the OpenAI API; the free ChatGPT plan does not include it.
OpenAI's launch announcement reports state-of-the-art scores on FrontierMath, ARC-AGI-3 and computer-use benchmarks. Those are the maker's own figures, so treat an independent leaderboard as the check on any specific claim.
API pricing is per token, with separate rates for short and long context, plus Batch, Flex and Fast-mode variants. Prompts over 272K input tokens are priced at the higher long-context rate for the whole request.
API use is priced per token and adds up fast; the free ChatGPT plan does not include Astra.
Published price
- Standard (≤272K input)$10 in / $50 out per 1M tokens
- Cached input$1.00 per 1M tokens
- Long context (>272K input)The full request is priced at the long-context rate$20 in / $75 out per 1M tokens
- Batch & Flex50% of standard rates
- Fast mode2× standard rates
- In ChatGPTAvailable to Plus, Pro, Business and Enterprise users, per OpenAI's announcementPlus $20, Pro $100/month
Source: OpenAI's API pricing page · Last checked 27 September 2026
Benchmarks
The published figures for GPT-6 Astra, copied exactly as their sources give them and checked 17 September 2026. Each one names the test, who ran it and when, and says when the number comes from the maker rather than an independent tester.
- Artificial Analysis Intelligence Index v4.3.253 — 7th of 216 models
combined score across ten reasoning and knowledge tests · tested: max reasoning
Artificial Analysis · 28 September 2026 · independent
Measured at 63.6 tokens per second output speed.
Figures copied from the sources above, checked 17 September 2026.
Where to read today's numbers
Scores move, so these are the boards that keep them current.
- LMArena leaderboardIndependent
Ranks models by blind head-to-head votes from the public. Good for a feel for general quality, weak on specialist work.
- Artificial AnalysisIndependent
Independent testing of quality, speed and price per million tokens across hosted models.
- Vals AI benchmarksIndependent
Independent evaluations on real professional tasks — law, finance, tax, medicine — rather than vendor demos.
How it performs, in my experience
OpenAI's flagship model for complex reasoning, coding and computer use. Its launch announcement reports state-of-the-art scores on several benchmarks; those are the maker's own figures, so check them against an independent leaderboard before relying on a specific claim.
Prompting it from this site
You can run this here
Available here as the “thinking” ChatGPT setting — it runs on this site's own AI credits, not your ChatGPT plan.
Use it for
- Demanding reasoning or research work
- Agentic computer-use tasks
- Complex coding through ChatGPT paid plans