> GPT-6 Astra is a closed AI model from OpenAI, released 4 Sep 2026. Index score 197 (GPT-4 = 100), rank 2 of 229 ranked models. API price $10 in / $50 out per million tokens, 1.05M context. [Models](https://themodelindex.org/models/) /[OpenAI](https://themodelindex.org/labs/openai/) # GPT-6 Astra OpenAI Released 4 Sep 2026 Closed Flagship [Compare](https://themodelindex.org/compare/gpt-6-astra-vs-kimi-k3/) [Announcement](https://openai.com/index/gpt-6-astra-next-generation-work/) 197186–209 Index score #2 of 229 models 5/ 8 Core tests taken 4 of 6 areas · plus 18 anchor tests Frontier Versus the frontier at release set or matched the record $10/ $50 Price per million tokens input / output, list price 1.05M Context window tokens it reads at once ## Where it sits · on the frontier _Chart: Model Index score of every scored model since late 2022, with the best closed and best open model over time_ Among the first models to reach this score. ## Test results · by area · tick = best by any model Show 15 anchor tests Reasoning & knowledge evidence GPQA Diamond 96% Humanity's Last Exam 55% Maths evidence FrontierMath T1-3 94% Coding evidence SWE-bench Verified no result Terminal-Bench no result Agents evidence APEX-Agents 65% Novel problems evidence ARC-AGI-2 95% Long tasks no evidence yet METR time horizon no result Independent results from the Epoch AI benchmarking hub, best reasoning setting. The core tests are shown here; anchor tests also inform the score and can be shown above. ## What others say · rank on each leaderboard The Model Index independent tests, our method #2 of 219 · 197 Epoch Capabilities Index overall capability from many benchmarks #2 of 239 · 166.5 LMArena · WebDev building web apps and front ends #2 of 123 · 1789 Artificial Analysis Intelligence Index overall capability across ten evaluations #4 of 207 · 53 LMArena · Coding coding questions in chat #9 of 383 · 1539 LMArena · Maths maths questions in chat #25 of 373 · 1490 LiveBench · Overall overall, with fresh questions to limit contamination #4 of 58 · 82.2 LMArena · Text overall chat quality, judged by people #27 of 388 · 1476 LMArena · Creative writing creative writing #37 of 386 · 1448 LMArena · Vision understanding images #18 of 149 · 1285 LMArena · Documents working with long documents #15 of 41 · 1468 Further left is better. Leaderboards measure different things: LMArena is people voting blind between two answers, the others are test scores. Click any row for the source. ## Best at · top three among current models 1st reasoning questions LiveBench · Reasoning 1st data analysis questions LiveBench · Data Analysis 1st GPQA Diamond Epoch AI · independent 1st Humanity's Last Exam Epoch AI · independent 1st FrontierMath T1-3 Epoch AI · independent 1st ARC-AGI-2 Epoch AI · independent 2nd building web apps and front ends LMArena · WebDev 3rd mathematics questions LiveBench · Mathematics 3rd language questions LiveBench · Language Where this model places in the top three of today's models, on category leaderboards and on our core tests. ## Where to use it · prices checked 1 Oct 2026 #### From the lab ChatGPT app · subscription ↗ OpenAI API API ↗ #### API providers · 3 · per million tokens, in / out O OpenAI $60 / $300 cheapest A Azure $11 / $55 ↗ A Amazon Bedrock $11 / $55 ↗ Provider prices come from OpenRouter's live listing; going direct to a provider can differ. Subscriptions are the lab's own apps. ## Facts and sources Released 4 Sep 2026 Announcement [openai.com ↗](https://openai.com/index/gpt-6-astra-next-generation-work/) Release entry GPT-6 Astra (preview 09-03) Scored as GPT-6 Astra Artificial Analysis index 53 · max Output speed 51 tokens/s Hugging Face closed weights ## OpenAI releases · around this one GPT-5.6 Sol Closed 9 JUL 2026 index 184 GPT-5.6 Terra Closed 9 JUL 2026 index 178 GPT-6 Astra Closed 4 SEP 2026 index 197 GPT-6 Luna Closed 22 SEP 2026 index 166 GPT-6 Sol Closed 22 SEP 2026 index 180 --- Source: https://themodelindex.org/models/gpt-6-astra/ · The Model Index · data as of 2026-10-01