What Is GPT-6 Astra?
The concise, source-backed explanation of OpenAI’s September 2026 flagship model.
GPT-6 Astra is OpenAI’s flagship model for difficult end-to-end work. It was announced on September 3, 2026 and is designed for complex reasoning, coding, research, computer use and document creation.
GPT-6 Astra key facts
What changed with Astra?
Astra combines a very large 1,050,000-token context window with up to 128,000 output tokens and a broad tool set. OpenAI’s model documentation lists web search, file search, image generation, code interpreter, hosted shell, apply patch, computer use, MCP and tool search among supported Responses API tools.
What can GPT-6 Astra accept and return?
The model accepts text and image input and returns text. Audio and video are not listed as native model modalities. Streaming, function calling and structured outputs are supported.
Who is Astra for?
Astra makes the most sense when model quality can justify premium token pricing: long-running research, complex coding changes, agentic workflows, document-heavy work and tasks that require multiple tools. For high-volume commodity generation, lower-cost models can be more economical.
Availability
OpenAI announced a staged rollout to ChatGPT Plus, Pro, Business and Enterprise users, along with API access and availability through Microsoft Azure and AWS Bedrock. Because rollout status can change account by account, treat the model picker or API account as the final availability check.
Frequently asked questions
When was GPT-6 Astra released?
OpenAI announced GPT-6 Astra on September 3, 2026.
Who makes GPT-6 Astra?
GPT-6 Astra is developed by OpenAI.
Does GPT-6 Astra support images?
Yes. OpenAI lists image input as supported, with text as the output modality.
Primary sources
Specifications and prices can change. We prefer first-party documentation and record the date each page was reviewed.