# GPT-6.2 Astra

> For the complete documentation index, see [llms.txt](/llms.txt). Markdown versions of documentation pages are available by appending `.md` to the page URL.

> Our most capable model for your hardest projects, from start to finish.

Model ID: `gpt-6.2-astra`

GPT-6.2 Astra can build and revise software, use a computer, and work with documents.
It accepts text and images and responds with text.
In the Responses API, use the [Multi-agent beta](/api/docs/guides/responses-multi-agent) to delegate work to subagents.

The API rejects requests that set reasoning effort to `none`.

The Responses API supports [discounted background completion windows](/api/docs/guides/background#use-discounted-completion-windows) in public beta.

## Model details

- Input modalities: text, image
- Output modalities: text
- Unsupported modalities: audio, video
- 1,050,000 context window
- Maximum input tokens: 922,000
- 128,000 max output tokens
- May 18, 2026 knowledge cutoff
- Reasoning token support

## Pricing

Prices are per 1M tokens. See the [pricing page](/api/docs/pricing) for processing tiers and billing details.

### Text tokens

| Metric | Price | Unit |
| --- | ---: | --- |
| Input | $8 | 1M tokens |
| Cached input | $0.4 | 1M tokens |
| Cache writes | $10 | 1M tokens |
| Output | $40 | 1M tokens |

- The cache write price is for 30 minutes of retention. A 12-hour cache write costs $20 per million tokens for short-context requests at Standard pricing.
- For requests with more than 272K input tokens, input and cache rates double and output costs 1.5 times the short-context rate for the full request.
- Batch costs 50% of Standard rates. Fast mode costs twice the Standard rate for the same context length.
- Ultrafast costs 3 times the Standard rate through February 28, 2027, and 6 times the Standard rate starting March 1, 2027 (Pacific time).

## Endpoints

| Endpoint | Route | Support |
| --- | --- | --- |
| Chat Completions | `v1/chat/completions` | Supported |
| Responses | `v1/responses` | Supported |
| Realtime | `v1/realtime` | Not supported |
| Assistants | `v1/assistants` | Not supported |
| Batch | `v1/batch` | Supported |
| Fine-tuning | `v1/fine-tuning` | Not supported |
| Image generation | `v1/images/generations` | Not supported |
| Videos | `v1/videos` | Not supported |
| Image edit | `v1/images/edits` | Not supported |
| Speech generation | `v1/audio/speech` | Not supported |
| Transcription | `v1/audio/transcriptions` | Not supported |
| Translation | `v1/audio/translations` | Not supported |

## Supported features

- streaming
- structured_outputs
- function_calling
- file_search
- image_input
- web_search
- prompt_caching

## Unsupported features

- fine_tuning
- predicted_outputs

## Supported tools

Tools supported by this model when using the Responses API.

- web_search
- file_search
- image_generation
- code_interpreter
- hosted_shell
- apply_patch
- skills
- computer_use
- mcp
- tool_search

## Model IDs

- `gpt-6.2-astra`

## Rate limits

Rate limits ensure fair and reliable access to the API by placing specific caps on requests, tokens, audio duration, or other usage within a given time period. Your usage tier determines how high these limits are set and automatically increases as you send more requests and spend more on the API.

### Standard

| Tier | RPM | TPM |
| --- | ---: | ---: |
| Build | 5,000 | 1,000,000 |
| Launch | 10,000 | 4,000,000 |
| Grow | 15,000 | 40,000,000 |
