benchrOpenAI
GPT-6 Astra
OpenAI · frontier. Long end-to-end work OpenAI aims this model at - reasoning, coding, computer use, research.
The published record
Every number below was read from the provider's own documentation on the date shown. benchr does not restate a figure it has not seen published.
- API identifier
gpt-6-astra- Context window
- 1050KtokensSource
- Maximum output
- 128KtokensSource
- Input
- $10.00per 1M tokensSource
- Output
- $50.00per 1M tokensSource
- Cached input
- $1.00per 1M tokensSource
- Released
- September 3, 2026Source
- License
- Proprietary
Availability Generally available since September 3, 2026. Listed on the official model directory as OpenAI's most capable model.
Record verified September 11, 2026
What the record says
Announced in OpenAI's API changelog on September 3, 2026 as "our most capable model, built for the hardest end-to-end work", aimed at reasoning, coding, computer use, research and document creation. The model directory lists a 1.05M context window and an April 30, 2026 knowledge cutoff; benchr recorded maximum output as null until OpenAI's model page, read on 2026-09-11, listed 128,000 max output tokens and a 922,000-token maximum input alongside the 1,050,000 context window. The same page lists reasoning effort levels low, medium, high, xhigh and max, and support for Chat Completions, Responses and Batch. The changelog records real interface constraints: the model does not support the `none` reasoning-effort level, does not accept custom temperature, top_p or logprobs, requires the Responses API for tool calling rather than Chat Completions, and is subject to misalignment monitoring. Those are migration blockers, not footnotes. benchr previously recorded that Astra was announced but unreleased with no model card, API id or price; that was accurate until this release and is superseded by this record.
What has moved
Entries in the change ledger that affect this model, newest first.
Released
OpenAI announced gpt-6-astra in its API changelog on September 3, 2026 as its most capable model, for reasoning, coding, computer use, research and document creation. The model directory lists a 1.05M context window and an April 30, 2026 knowledge cutoff, and publishes no maximum-output figure. Pricing is $10.00 input / $1.00 cached input / $50.00 output per 1M tokens, with Batch at half of each. The changelog records interface constraints that are migration blockers: no `none` reasoning-effort level, no custom temperature, top_p or logprobs, and tool calling only through the Responses API. Source: developers.openai.com/api/docs/changelog, /pricing and /models (verified 2026-09-08).
Which one to use
The rest of the family, with the two numbers that usually decide it. The current model is marked.
| Models | Input | Output | Context window |
|---|---|---|---|
| GPT-5 Mini | $0.25 | $2.00 | 400K |
| GPT-5 | $1.25 | $10.00 | 400K |
| GPT-5.4 | $2.50 | $15.00 | 1050K |
| GPT-5.6 | $4.00 | $20.00 | 1050K |
| GPT-5.5 | $5.00 | $30.00 | 1050K |
| GPT-6 Astra This page | $10.00 | $50.00 | 1050K |
benchr's read
Worth it for
- Long end-to-end work OpenAI aims this model at - reasoning, coding, computer use, research
- Prompts that need more than the 400K window of the GPT-5.6 family
- Teams already on the Responses API, which this model requires for tool calling
Look elsewhere if
- You send custom temperature, top_p or logprobs - this model accepts none of them
- You rely on the `none` reasoning-effort level, which it does not support
- You call tools through Chat Completions rather than the Responses API
- Cost is the constraint: input is 2.5x GPT-5.6 Sol and output is 2.5x its rate
- You need benchr-held benchmark figures - OpenAI published none with this release
Prices, limits, and identifiers above are provider facts. This section is benchr's judgement about them.
What benchr has written about it
Pieces that name this model, newest first. The ones written about this model come before the ones that mention it in passing.
What is not on this page
Stated rather than filled in.
- No first-token or tokens-per-second figure is recorded. benchr has not measured it and the provider does not publish one.