OpenAI · Model guide
GPT-6.2 Astra
OpenAI's most capable GPT-6.2 model, built for the hardest projects from start to finish: it builds and revises software, uses a computer and works with documents. It accepts text and images, supports a 1.05M-token context window, and always reasons: reasoning effort none is not accepted.
What this model is good at
What the vendor positions it for, and which of our models to reach for instead.
VendorOur most capable model for your hardest projects, from start to finish.source
Software engineering · Build and revise software, including demanding coding tasks.(vendor claim)source
Computer use · Operate a computer to carry out tasks.(vendor claim)source
Working with documents · Create, read and revise documents.(vendor claim)source
Demanding reasoning · The most demanding reasoning work in the GPT-6.2 family.(vendor claim)source
- openai/gpt-6.2-solcheaper —Use GPT-6.2 Sol for everyday and complex work at lower cost.
- openai/gpt-6-astrapredecessor —Use GPT-6 Astra to keep an established baseline.
Pricing and billing
Tiered pricing: the rate changes once the input passes the threshold.
| Price / 1M tokens by input size | Input < 272.0K | Input ≥ 272.0K |
|---|---|---|
| Input | $ 8.00 | $ 16.00 |
| Output | $ 40.00 | $ 60.00 |
| Cache read | $ 0.40 | $ 0.80 |
| Cache write | $ 10.00 | $ 20.00 |
At this input size you are on the Input < 272.0K tier.
+ 2.0K × $ 40.00
Capabilities and limits
What CrossModel guarantees across every route this model can take right now.
- Context window
- 1.1M tokens
- Max output
- 128.0K tokens
- Input / output modalities
- Text + Image → Text
- Streaming
- Supported
- Tool calling
- Supported
- Structured output (JSON)
- Supported
- Image input
- Supported
- Thinking mode
- Supported· always on
- Reasoning effort
- low · medium · high · xhigh
- Available endpoints
- /v1/chat/completions · /v1/responses · /v1/messages
- Tool calling is available only when source
Published benchmarks
Scores the vendor reported, with the evaluation setup each one came from.
- ExploitBenchscore·OpenAI · 2026-11-17setup not fully disclosed
System card 2.1.2.1.2 节(Table 4)ExploitBench V8 的 mean capability credit(41 个漏洞、每个 5 次尝试)。OpenAI 说明 GPT-6.2 Astra 在每个推理档位都拿到 100%;历史漏洞(最早到 2023 年)有训练数据污染风险,成绩可能虚高。同表 GPT-6 Astra 100%、GPT-6.2 Sol 100%、GPT-6.1 Sol 99.7%。研究/API 配置可能不同于生产。
100%
Vendor-reported numbers, not CrossModel measurements. Scores are only comparable when the benchmark version, metric and evaluation setup match, so nothing here is averaged or ranked.
Use it in your tools
Point the base URL at CrossModel and paste this model ID — every tool below has a setup guide.
Frequently asked questions
What is GPT-6.2 Astra?
How much does GPT-6.2 Astra cost?
Does GPT-6.2 Astra support tool calling and structured output?
Which endpoint do I call?
Can I try it without writing code?
Chat first, integrate later
Chat first, then wire it in once you like the answers.