gpt-6-astra

gpt-6-astra is OpenAI’s flagship model for the most difficult end-to-end tasks. The official model catalog describes it as the highest competency level suitable for computer use, browsing, software engineering, science, and professional work; it supports multiple levels of inference intensity and a variety of tools, including web search, file search, computer use, function calls, and MCP. It is suitable for multi-step task orchestration, complex code base changes and professional processes that require strict boundary control. When using it, the executable operations, approval points, and stop conditions should be clearly written into the instructions.

Context

1050000 tokens

Max output

128000 tokens

Input

llm image

Output

llm