GPT-6.2 Astra
GPT-6.2 Astra is OpenAI's frontier model for demanding professional work, complex coding, computer use, scientific workflows, and polished documents.
Overview
Released November 17, 2026. GPT-6.2 Astra accepts text and image input, has a 1.05M-token context window, a 922K maximum input, and up to 128K output tokens. It is designed for complex professional work, coding, computer use, science, and document creation. Tool calling and hosted tools require the Responses API; reasoning effort none is unsupported. It is available in ChatGPT Work, Codex, and the OpenAI API.
Model specifications
- Family
- GPT-6
- Context
- 1,050,000 tokens
- Max input
- 922,000 tokens
- Max output
- 128,000 tokens
- Input
- Text, Image
- Output
- Text
- Open weights
- No
- Capabilities
- Reasoning, Vision, Tool Use, Function Calling, Structured Output, Web Search, Code Execution, Computer Use, Image Generation
Recorded API token pricing
Reference rates in USD per 1 million tokens. Actual charges depend on the provider, service tier, context length, tools, and region.
- Input
- USD 8 / 1M tokens
- Cached input
- USD 0.4 / 1M tokens
- Output
- USD 40 / 1M tokens
Pricing caveats
Standard short-context USD rates per 1M text tokens. Cache writes: $10 (30m) / $20 (12h). Above 272K input tokens: 2x input/cache and 1.5x output for the full request. Batch/Flex 0.5x; Fast 2x; Ultrafast 3x through Feb 28, 2027, then 6x. Eligible 2h/8h background completions: 30%/40% below Standard. Tool calls may be billed separately.
Reviewed Nov 18, 2026 · Verify with provider documentation →