Skip to content

AI Agents & Coding · Paid

GPT-6.2 Astra

GPT-6.2 Astra is OpenAI's frontier reasoning model, released on November 17, 2026 and available in the OpenAI API under the model ID gpt-6.2-astra. OpenAI describes it as its most capable model for the hardest projects, aimed at complex coding, computer use, research and professional documents such as slides, spreadsheets and legal or financial work. It replaces GPT-6 Astra at the top of OpenAI's lineup.

Who it is for: developers and teams running demanding agentic work, long-horizon coding, scientific or analytical tasks, and document generation where quality matters more than cost per request. For everyday coding and high-volume agent work, OpenAI points developers to the cheaper GPT-6.2 Sol.

Key features, per OpenAI's model page (accessed November 18, 2026): - Context window of 1,050,000 tokens, up to 922,000 input tokens and 128,000 output tokens; knowledge cutoff May 18, 2026. - Text and image input, text output; reasoning tokens. The API rejects a reasoning effort of none. - Works with the Responses API and Chat Completions, but tool calling requires the Responses API, where it supports web search, file search, image generation, code interpreter, hosted shell, apply patch, skills, computer use, MCP and tool search, plus the Multi-agent beta. - Batch API support; no fine-tuning or predicted outputs. - Also generally available on Amazon Bedrock and in GitHub Copilot (Pro+, Max, Business and Enterprise) from launch day.

Pricing type: paid, per token. OpenAI's Standard rates are $8 per million input tokens, $0.40 per million cached input tokens, $10 per million tokens for 30-minute cache writes and $40 per million output tokens, which is 20% below GPT-6 Astra on input and output and 60% below it on cached input. Requests over 272K input tokens are billed at double the input and cache rates and 1.5 times the output rate for the whole request. Batch costs 50% of Standard, Fast mode twice Standard, and Ultrafast three times Standard through February 28, 2027, then six times. In the Responses API public beta, two-hour and eight-hour background completion windows cut token rates by 30% and 40%.

Benchmarks: Artificial Analysis lists GPT-6.2 Astra (Max) at 62 on its Intelligence Index v5.0.1, at about $4.00 per index task (accessed November 18, 2026). OpenAI's own launch results are vendor-reported.

Alternatives: GPT-6.2 Sol and GPT-6 Astra from OpenAI, Claude Fable 5.2 and Claude Opus 5.5 from Anthropic, and Gemini 4 Carbon from Google.

Official documentation: https://developers.openai.com/api/docs/models/gpt-6.2-astra

Official site