model releaseOpenAI

OpenAI releases GPT-6.2 Astra with 1.1M context at $8 input, $40 output per 1M tokens

TL;DR

OpenAI released GPT-6.2 Astra on Nov 17, 2026, listing a 1.1M-token context window and pricing of $8 input and $40 output per 1M tokens. OpenAI reports new highs on its evals at less than a third of the cost per task of GPT-6 Astra.

1 min read
0

GPT-6.2 Astra at a glance

Context window1050K tokens
Input$8/1M tokens
Output$40/1M tokens

How GPT-6.2 Astra ranks now

BenchmarkScoreRankSource
AA Intelligence Index62.1 index2 of 54Artificial Analysis, Nov 2026 ↗
Speed (tok/s)64 tok/s77 of 103Artificial Analysis, Nov 2026 ↗

Rank is against every model in our database with a sourced score on that benchmark, each counted once with its best score.

All GPT-6.2 Astra specs and scores →

OpenAI released GPT-6.2 Astra on Nov 17, 2026, a flagship model in the GPT-6.2 series with a 1.1M-token context window. It is priced at $8 per 1M input tokens and $40 per 1M output tokens, according to its OpenRouter listing (model ID openai/gpt-6.2-astra).

Key specifications

  • Model: GPT-6.2 Astra (OpenAI)
  • Context window: 1.1M tokens
  • Input price: $8 per 1M tokens
  • Output price: $40 per 1M tokens
  • Release date: Nov 17, 2026
  • Parameter count, training cutoff, benchmark scores: not yet disclosed

Positioning and capabilities

OpenAI describes GPT-6.2 Astra as the flagship model in the GPT-6.2 series, built for the hardest projects. The listing says it is suited for real-world engineering, professional work and scientific research. It also cites stronger visual reasoning for websites, slideshows and 3D designs.

Cost claim

OpenAI reports new highs on its evals at less than a third of the cost per task of GPT-6 Astra. This is a company claim, and specific eval names and scores have not yet been published in the listing.

Availability

As of Nov 17, 2026, the OpenRouter listing states that GPT-6.2 Astra is not currently available through EU in-region routing. The model is accessible through the API.

What this means

GPT-6.2 Astra pairs a 1.1M-token context window with a stated focus on engineering, professional and scientific work, and on visual tasks such as websites, slideshows and 3D designs. The cost-per-task claim relative to GPT-6 Astra is OpenAI's own, so teams should test it on their own workloads rather than rely on the headline figure. Organizations that require EU in-region routing cannot use the model that way as of the listing date.

Related Articles

Comments

Loading...