Skip to content

ModelsOpenAIUpdated

OpenAI launches GPT-6.2 Astra: frontier model for professional work at 20% lower API prices

GPT-6.2 Astra at a glance

OpenAI launched GPT-6.2 Astra on November 17, 2026, in ChatGPT Work, Codex and the API (gpt-6.2-astra) at $8 input and $40 output per million tokens. As of November 17, 2026, Artificial Analysis ranks it #2 of 111 models on its Intelligence Index, at 62 with max effort.

Current standing

As of

Rank
#2of 49 models

GPT-6.2 Astra (Max) · 62.1 on the Artificial Analysis Intelligence Index v5.0.1 (opens in a new tab)

Recorded at publication
62 (v5.0.1)

as of

Current leader
Claude Fable 5.2 (Anthropic) · 64.6

Computed from our daily snapshot; it changes as new models are measured. OpenAI launches GPT-6.2 Astra: frontier model for professional work at 20% lower API prices

Ask your AI about this story

Ask what changed in GPT-6.2 Astra and what to check before using it in your business. Compare it on price, speed, intelligence and the price-to-speed ratio against the models I would otherwise use, and say which tasks it suits or does not suit.

What changed with GPT-6.2 Astra

On November 17, 2026, OpenAI released GPT-6.2 Astra for complex professional work. In ChatGPT Work and Codex, it is available on Plus, Pro, Enterprise and Edu, with eligible Enterprise and Edu access enabled by the workspace admin. The API model gpt-6.2-astra takes text and images and returns text. It supports a 1,050,000-token context window, with up to 922,000 input and 128,000 output tokens.

GPT-6.2 Astra computer-use example in OpenAI's announcement: a 3D inline-six engine simulation at 8,000 RPM showing gas-temperature and mechanical-stress heat maps side by side

OpenAI uses this engine simulation to illustrate the level of detail Astra can work with in CAD and simulation tools.

In the November 17, 2026 Artificial Analysis snapshot, GPT-6.2 Astra scored 62 at max effort and placed #2 of 111 models. In its own release evaluation, OpenAI reported that Astra scored 5.2 points above GPT-6 Astra on Agents' Last Exam at less than one-third of the cost per task. These are results for named evaluations. Standard API pricing per million tokens is $8 input, $0.40 cached input and $40 output. OpenAI documents higher rates for requests above 272,000 input tokens.

Scatter plot for Agents' Last Exam v1.1. The horizontal axis shows cost per task and the vertical axis shows benchmark score. The labeled points include GPT-6.2 Astra, GPT-6.2 Sol, GPT-6 Astra and Claude Fable 5.2.

In its November 17 release, OpenAI reported that GPT-6.2 Astra scored 5.2 points above GPT-6 Astra on Agents' Last Exam at less than one-third of the cost per task.

OpenAI showed Astra following a slide template and producing workbooks and memos from source documents. The announcement also shows mechanical design in CAD and a side-by-side 3D garden demo. These examples give teams concrete outputs to evaluate, from formatting fidelity to how well a model handles requested revisions.

GPT-6.2 Astra 3D design comparison in OpenAI's announcement: GPT-6 Astra's flat, low-detail 3D garden with raised beds beside GPT-6.2 Astra's realistic version with wood grain, a trellis and a tomato plant

OpenAI shows the earlier Astra result and the GPT-6.2 Astra result side by side for the same 3D garden example.

GPT-6.2 Astra benchmarks and pricing

Values are copied from the linked source on the date shown. They are not IA Compa measurements.

Artificial Analysis Intelligence Index (max)as of · Source for Artificial Analysis Intelligence Index (max): artificialanalysis.ai (opens in a new tab)
62 (v5.0.1)
Output speed (max)as of · Source for Output speed (max): artificialanalysis.ai (opens in a new tab)
66 tokens/s
Output price per 1M tokensas of · Source for Output price per 1M tokens: developers.openai.com (opens in a new tab)
$40.00
Context windowas of · Source for Context window: developers.openai.com (opens in a new tab)
1.05M tokens
Cached input price per 1M tokensas of · Source for Cached input price per 1M tokens: developers.openai.com (opens in a new tab)
$0.40

How it compares

Current models we have covered: released in the last six months, newest version of each line only. Each value is the model's latest recorded value. (Models)

Intelligence indexhigher is better
Output speed (tokens/s)higher is better
Input price (USD per 1M tokens)lower is better
Output price (USD per 1M tokens)lower is better
Context window (tokens)higher is better

GPT-6.2 Astra reasoning effort

Artificial Analysis tests this model at several reasoning-effort settings. Higher effort usually scores higher but costs more and answers more slowly.

GPT-6.2 Astra by effort level

Higher effort lets the model think longer. Here is what each setting scores, how fast it writes, and what it costs on Artificial Analysis.

Intelligence Index · higher is better
Cost to run the index · lower is betterHover, tap or focus a point for exact numbers.
GPT-6.2 Astra: Artificial Analysis results by reasoning effort
EffortIntelligence IndexOutput speed (tokens per second)Cost to run the indexBlended price (3:1) (per 1M tokens)Price in / out (per 1M tokens)First token
Max (GPT-6.2 Astra (max)) · Best: highest effort with a score, used in the leaderboard6266$4,530$16$8 / $40135 s
Extra high (GPT-6.2 Astra (xhigh))6263$2,484$16$8 / $4036 s
High (GPT-6.2 Astra (high))6062$2,007$16$8 / $4015 s
Medium (GPT-6.2 Astra (medium))5760$1,457$16$8 / $402.7 s
Low (GPT-6.2 Astra (low))4660$794$16$8 / $402.4 s
Price: $8 / $40 per 1M input / output tokens at every effort.Artificial Analysis · as of Nov 17, 2026 (opens in a new tab)

Benchmarked by AA: max · xhigh · high · medium · low of 6 available

Where to compare GPT-6.2 Astra

Independent leaderboards and official resources for comparing this model yourself. Rankings change often; each site shows its own current data.

Why GPT-6.2 Astra matters for businesses

Why it matters: Try Astra first on work that takes several rounds of review, such as a board deck, a workbook with assumptions to check, or a large code change. Compare output quality and cost per completed task with GPT-6.2 Sol before changing the model for the rest of your work.

Put it to work: AI services

Get AI News by email

Verified AI stories, each with its official source, on Monday and Thursday mornings. Add breaking news to hear about the biggest ones as soon as we publish them.

Every story links its official source

Free, in English or Spanish

Unsubscribe with one click, any time

What to send

GPT-6.2 Astra updates since publication

Dated changes since publication. The original article text above is kept as written on its date.

  1. Update (Updated)

    Update, November 17, 2026: Artificial Analysis has published measured performance for GPT-6.2 Astra: at max effort it averages 10.4 minutes per Intelligence Index task and 134.86 seconds to the first answer token. Astra leads AA-3D at 1318, matches GPT-6 Astra on the Coding Agent Index, and its AA-Omniscience hallucination rate at max effort is 59%.

    artificialanalysis.ai (opens in a new tab)

Sources for this story

  1. Official
  2. GPT-6.2 Astra model page (opens in a new tab)

    OpenAI · developers.openai.com

    Official
  3. Official
  4. GPT-6.2 system card (opens in a new tab)

    OpenAI · deploymentsafety.openai.com

    Official
  5. GPT-6.2 Astra model page (opens in a new tab)

    Artificial Analysis · artificialanalysis.ai

    Benchmark
  6. Artificial Analysis LLM Leaderboard (opens in a new tab)

    Artificial Analysis · artificialanalysis.ai

    Benchmark

Last verified: . IA Compa summarizes public announcements for business readers. We are not affiliated with the companies mentioned. Always confirm details, pricing, and availability with the official source.

  • #llm
  • #openai
  • #gpt-6
  • #gpt-6-2
  • #frontier-model
  • #agents
  • #codex

Want this working in your business?

We turn new AI capabilities into agents, voice receptionists, and automations you can actually use.