GPT-6 Astra,
decoded.
A source-first guide to OpenAI’s newest flagship model — specs, API pricing, long-context rules, availability, comparisons and practical implementation.
- Context window
- 1.05M tokens
- Max output
- 128K tokens
- Input / output · per MTok
- $10 / $50
- Knowledge cutoff
- Apr 2026
Beyond the launch.
New integrations, candid user discussions and things people are actually making.
Where Astra goes next.
Copilot, Vercel AI Gateway, Cline and the week’s research coverage.
Explore news →COMMUNITY / 4 STORIESWhat builders are saying.
Frontend experiments, token budgets and real integration friction.
Explore community →SHOWCASE / 3 STORIESMade with Astra.
A live landing page, a Blender remake and an action RPG.
Explore showcase →See Astra in action.
OpenAI’s launch film shows Astra moving beyond chat to handle computer tasks, software work and longer, multi-step workflows.
GPT-6 Astra at a glance
What is GPT-6 Astra?
GPT-6 Astra is OpenAI’s flagship model for its hardest end-to-end work. OpenAI announced Astra on September 3, 2026 and positions it for complex reasoning, coding, computer use, research and document creation.
For developers, the model ID is gpt-6-astra. It accepts text and image input, supports reasoning controls, streaming, function calling and structured outputs, and can use tools such as web search, file search, code interpreter and computer use through supported APIs.
Estimate GPT-6 Astra API cost
Enter token usage for a single request. Calculate separate requests individually; do not combine their token totals.
OpenAI states that prompts above 272K input tokens are billed at 2× input/cache rates and 1.5× output rates for the full request.
Estimate excludes tool-call fees and other service-specific charges.
Start with the question you have
Pricing
Rates, long-context rules and a live cost calculator
Explore →02API
Model ID, endpoints, tools and integration notes
Explore →03Context window
What 1.05M tokens means — and what >272K costs
Explore →04Benchmarks
Official results, filterable data and downloadable CSV
Explore →05Coding
Where Astra fits in software engineering workflows
Explore →06Comparisons
Astra vs GPT-5.6 Sol, Claude Opus 5 and Gemini 3.8 Flash
Explore →Track rollout, speed and benchmark data.
Facts first. Tests labeled. Changes logged.
Specifications are checked against primary documentation, vendor benchmarks stay labeled as vendor results, and future independent tests follow a published methodology.
Astra rollout timeline
GPT-6 Astra announced.
OpenAI begins a staged rollout and publishes core model specifications.
Model documentation verified.
This guide confirms API rates, 1.05M context, 128K output, supported features and long-context pricing rules against first-party docs.
Independent tests.
Latency, coding, long-context and tool-use tests will be added with prompts, settings and dates.
Frequently asked questions
What is GPT-6 Astra?
GPT-6 Astra is OpenAI’s flagship model for complex end-to-end work including reasoning, coding, computer use, research and document creation.
How much does GPT-6 Astra cost in the API?
OpenAI lists standard pricing at $10 per million input tokens, $1 per million cached input tokens and $50 per million output tokens. Prompts above 272K input tokens use higher long-context rates.
What is the GPT-6 Astra context window?
GPT-6 Astra has a 1,050,000-token context window and supports up to 128,000 output tokens.
What is the GPT-6 Astra API model ID?
The model identifier is gpt-6-astra.
Is GPT-6 Astra available in ChatGPT?
OpenAI announced a staged rollout to Plus, Pro, Business and Enterprise plans. Account availability can vary during rollout.