GPT-6.2 Astra
openai/gpt-6.2-astra
Back to model catalog

OpenAI · Model guide

GPT-6.2 Astra

Compare
openai/gpt-6.2-astraReleased Nov 17, 2026

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.

Modalities
TextImageText
Context
1M
Max output
128K
Price / 1M tokens
$ 8.00in$ 40.00out
StreamingTool callingJSON outputImage inputThinking mode

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

Best for
  • 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

Neighbouring models
Compare them side by side →

Pricing and billing

Tiered pricing: the rate changes once the input passes the threshold.

Price / 1M tokens by input sizeInput < 272.0KInput ≥ 272.0K
Input$ 8.00$ 16.00
Output$ 40.00$ 60.00
Cache read$ 0.40$ 0.80
Cache write$ 10.00$ 20.00
Estimate a request

At this input size you are on the Input < 272.0K tier.

$ 0.1600
Estimated cost per request
10.0K × $ 8.00
+ 2.0K × $ 40.00
≈ $ 160.00 per 1,000 requests

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
Conditional limitsDisclosed by the vendor. These do not make the capability unavailable — they say when it is not.
  • Tool calling is available only when source

Published benchmarks

Scores the vendor reported, with the evaluation setup each one came from.

Cybersecurity
  • ExploitBench
    score·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%
Knowledge work
  • HealthBench Professional
    score·OpenAI · 2026-11-17setup not fully disclosed

    System card 3.4.1 节(Table 17)的 length-adjusted 分数;未做长度调整为 66.0,平均回答长度 1884 字符。同表 GPT-6.2 Sol 65.9、GPT-6.1 Sol 64.2、GPT-6 Astra 64.7。该节未披露推理档位;研究/API 配置可能不同于生产。

    66.3%
Scientific reasoning
  • GPQA Diamond
    accuracy·OpenAI · 2026-11-17setup not fully disclosed

    System card 2.1.1.1 节(Table 2)GPQA Diamond (no tools) 准确率,作为生物能力评测的一部分报告。未披露推理档位;评测在无生产安全防护下进行。同表 GPT-6.2 Sol 96.0%、GPT-6 Astra 96.0%、GPT-6.1 Terra 93.56%。研究/API 配置可能不同于生产。

    97%

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.

See all integration guides →

Frequently asked questions

What is GPT-6.2 Astra?
GPT-6.2 Astra is available on CrossModel as openai/gpt-6.2-astra. It has a 1M-token context window and can return up to 128K tokens per request.
How much does GPT-6.2 Astra cost?
Currently $ 8.00 per 1M input tokens and $ 40.00 per 1M output tokens. Cached input is billed at $ 0.40 per 1M.
Does GPT-6.2 Astra support tool calling and structured output?
Tool calling: Supported · Structured output (JSON): Supported · Image input: Supported · Streaming: Supported · Thinking mode: Supported
Which endpoint do I call?
For text models, OpenAI Chat Completions (/v1/chat/completions), OpenAI Responses (/v1/responses) and Anthropic Messages (/v1/messages) all work — pick whichever your SDK already speaks.
Can I try it without writing code?
Yes. Sign in and open the Playground on this page to chat with the model directly — no API key needed. Usage is billed from your wallet at the normal rate.

Chat first, integrate later

Chat first, then wire it in once you like the answers.