Skip to main content
AI Choice EngineAI Choice Engine

OpenAI

GPT-6 Astra

GPT-6 Astra is a OpenAI model with 1.05M context, $10 in / $50 out per million tokens, last verified September 5, 2026. Open weights: no. Public-board scores are linked to the publisher; vendor-only figures stay in the watch-out, not on the board.

GPT-6 Astra is OpenAI's computer-use flagship — the first model it gates at the Preparedness Framework's 'Critical' cybersecurity threshold, trained on >100,000 GPUs at Stargate.

Catalog record checked September 5, 2026Individual provider fields may change

AI model specification details
SpecificationGPT-6 Astra
ProviderOpenAI
TierFrontier
Context window1.05M
Max output128K
Input / 1M tokens$10
Output / 1M tokens$50
WeightsClosed
Parameterscomputer-use flagship
Reasoning levelslow, medium, high, xhigh, max
Modalitiestext, image
API model idgpt-6-astra
ReleasedSeptember 3, 2026

Pricing tiers: Standard $10/$50 per MTok (≤272K context); >272K input billed $20/$75 (Fast mode output $100). Cached input $1/MTok; Fast mode is 2x price for ~2.5x speed. Knowledge cutoff Apr 30 2026. Rolling out over several days from 2026-09-03; Enterprise tenants get it disabled by default.

Verified evidence

Published benchmark results

Each result keeps its source and measurement date visible. A missing benchmark is not treated as a zero.

Published benchmark results for GPT-6 Astra
BenchmarkScoreMeasuredSource
Artificial Analysis Intelligence Index [max]61.22026-09-03Artificial Analysis · View source
DeepSWE 1.174.12026-09-03OpenAI · View source
OSWorld 2.0 (offline)72.62026-09-03OpenAI · View source
GPQA Diamond962026-09-03OpenAI GPT-6 Astra launch chart (vendor; transcribed by Vellum) · View source
Humanity's Last Exam57.22026-09-03OpenAI GPT-6 Astra launch chart (vendor, with tools; transcribed by Vellum) · View source
Terminal-Bench 2.187.42026-09-03Terminal-Bench official leaderboard (independent, native harness, high effort; captured 2026-09-03 — the live board has since moved to Terminal-Bench 4.0) · View source
Terminal-Bench 2.189.92026-09-05Artificial Analysis Terminal-Bench v2.1 (independent, Terminus 2 harness — conflicts with tbench.ai 87.4; harness difference preserved) · View source

Best for

  • Computer-use agents
  • Agentic coding
  • Hard reasoning

Watch out

Gated at the 'Critical' cyber threshold and disabled by default for Enterprise; long-context work bills $20/$75 above 272K input tokens.

History

GPT-6 Astra timeline

Release and status events for this model that passed two-source verification — dated, sourced, and linked.

  • GPT-6 AstraReleaseSeptember 3, 2026

    OpenAI launches GPT-6 Astra, its computer-use flagship and the first model gated at the Preparedness Framework's 'Critical' cybersecurity threshold — $10/$50 per MTok with a 1.05M-token context, 61.2 on the AA Intelligence Index and a self-reported 74.1% on DeepSWE 1.1.

    OpenAI — Introducing GPT-6 Astra

Source receipts

Catalog figures for this model were checked against the following sources.

Common questions

GPT-6 Astra

Answered from the verified figures on this page rather than general guidance.

How much does GPT-6 Astra cost per million tokens?

GPT-6 Astra is listed at $10 per million input tokens and $50 per million output tokens at standard rates. Output tokens usually dominate real bills, so weigh the output rate more heavily than the input rate. Standard $10/$50 per MTok (≤272K context); >272K input billed $20/$75 (Fast mode output $100). Cached input $1/MTok; Fast mode is 2x price for ~2.5x speed. Knowledge cutoff Apr 30 2026. Rolling out over several days from 2026-09-03; Enterprise tenants get it disabled by default.

What is GPT-6 Astra's context window?

GPT-6 Astra accepts about 1.05M tokens of context. That only matters if you routinely send very long documents, large codebases, or multi-turn histories that approach that limit.

What is GPT-6 Astra best for?

GPT-6 Astra is a frontier tier from OpenAI. It suits computer-use agents, agentic coding, hard reasoning. Gated at the 'Critical' cyber threshold and disabled by default for Enterprise; long-context work bills $20/$75 above 272K input tokens.