ai/list101

GPT-6.2 Astra

openai.com

OpenAI reasoning model for complex coding, scientific research, computer use and professional document workflows.

Llm Models
Visit website
Pricing
Paid
Available on
OpenAI API, ChatGPT Work, Codex
Category
Llm Models

What is GPT-6.2 Astra?

GPT-6.2 Astra is OpenAI's reasoning model for demanding technical and professional work. OpenAI positions it for software development, scientific research, computer interaction and work on documents, spreadsheets and presentations.

It accepts text and images and returns text. Developers can integrate it through the API, while eligible subscribers can select it in ChatGPT Work and Codex. Its focus makes it relevant to developers, researchers and professionals handling complex projects.

Key features

Software and frontend development

Build and revise software, including frontend implementations, using code and visual requirements.

Professional document work

Work on slides, spreadsheets and documents, including matching supplied presentation templates.

Connected tools

Responses API integrations support web search, file search, code interpreter, hosted shell, patch application, computer use and MCP.

Large context

The API documents a 1,050,000-token context window, with up to 922,000 input tokens and 128,000 output tokens.

Application integration

Streaming, function calling and structured outputs support interactive applications and machine-readable responses.

Who is it for?

Developers investigating complex bugs

Provide relevant code and failure details to request a diagnosis and proposed changes.

Researchers reviewing technical material

Supply research material and ask for an explanation of assumptions, methods and unresolved questions.

Professionals preparing deliverables

Provide source material and a template to request a presentation or spreadsheet draft.

How to use GPT-6.2 Astra

  1. Choose ChatGPT Work, Codex or an OpenAI API integration.
  2. For Work or Codex, update the app or CLI and select GPT-6.2 Astra when available.
  3. For API development, follow the official quickstart and use the model identifier gpt-6.2-astra.
  4. Provide the task and relevant text or images; configure any required tools.
  5. Review the response or proposed changes before using the result.

GPT-6.2 Astra pricing

Paid

API usage is metered by tokens. Eligible ChatGPT subscriptions provide Work and Codex access under their usage allowances. Processing tiers affect rates, and tools and cache writes can add charges.

Standard API, short context

USD 8 input, USD 0.40 cached input and USD 40 output per million tokens
Metered token consumption rather than a monthly model subscription

Applies to requests with up to 272,000 input tokens. Above that threshold, input and cache rates double and output rates increase by 50% for the entire request. Tools and cache writes can incur additional charges.

Check current pricing ↗

What to know before you use it

  • Native audio and video input are unsupported; the model generates text responses.
  • Fine-tuning is unsupported, and API requests cannot set reasoning effort to none.
  • Work and Codex availability depends on plan, workspace permissions and rollout. This release does not add Astra to regular Chat.

Frequently asked questions

Can Astra work with images?

Yes. It accepts images alongside text and generates text responses.

Where do subscribers select Astra?

In ChatGPT Work or Codex when their account has access. Enterprise and Edu access requires administrator enablement.

Does the model automatically include computer and search tools?

The model supports these tools through the Responses API, but an application must configure the corresponding integrations.