Skip to main content
OpenAIFlagship

GPT-6 Astra

OpenAI's most capable model. It researches, builds and checks its own work.

Build with GPT Astra

Each clip is one prompt, run on GPT Astra in Runable.

GPT Astra on Runable

  • Made for computer use

    It browses, fills forms and tests what it built.

  • Big, demanding jobs

    Full sites, deep research and complex documents.

  • Compare in one place

    Run the same task on Astra and Opus 5.5.

How to use GPT Astra on Runable

  1. Select GPT Astra

    Start a task on Runable and pick GPT Astra in the model menu.

  2. Describe the job

    Type what you need, or add files and links. Runable plans it and gets to work.

  3. Edit and ship

    Refine by chatting, then publish the site or export to PPTX, PDF or Google Slides.

One prompt, finished work

Real results from single Runable tasks. Pick one to watch it.

Made on Runable with GPT Astra
Build an award-winning single-page site for an environmental nonprofit called Verdant: an animated 3D Earth hero, live stats, and scroll-triggered animations.
token context window
1.05M
ScreenSpot-Pro, computer use
92.7%
BrowseComp, web research
91.5%
turn to build
Your

Why choose GPT Astra

  • Best-in-class computer use

    72.6% on OSWorld 2.0, ahead of Claude Opus 5.

  • Web research

    91.5% on BrowseComp, OpenAI’s hard-to-find-facts test.

  • Science and maths

    97.6% on FrontierMath Tier 4, 64.6% on Terminal-Bench Science.

  • Fewer mistakes

    4.2% hallucination rate, down from 12.2% on GPT-5.6 Sol.

  • Concise

    Fewer output tokens than most models, per Artificial Analysis.

  • Long context that holds

    96.3% on OpenAI’s MRCR long-context test.

GPT Astra specs

Maker
OpenAI
Family
GPT-6 (flagship)
Released
September 3, 2026
API model ID
gpt-6-astra
Context window
1.05M tokens
Max output
128K tokens
Input / output
Text and images in, text out
Reasoning
Effort from low to max
Knowledge cutoff
April 30, 2026
API price
$10 input / $50 output per 1M tokens
API prices are OpenAI’s list prices. On Runable you use your plan’s credits instead of paying per token. Runable’s picker shows this model as “GPT Astra”.

The only AI agent you need

Need SSO, data localization, premium support, onboarding assistance? Contact Sales

Put GPT Astra to work

Describe the job. Runable runs it on GPT Astra and ships the result.