OpenAI · Benchmarks

GPT-6 Astra benchmarks

GPT-6 Astra has 7 published scores in this dataset, from a OpenAI release dated 2026-09-03. Its composite index is 91.4, which places it 4th of 22 ranked models. Its strongest area is knowledge (index 99.1) and its weakest is coding (index 83.7). Its best placing is on GPQA Diamond, where its 96.1 ranks 1st of 19 models with a published score.

Closed APIOfficial page

Category profile

Index 0–100: each score relative to the best published score on that test.

ReasoningCodingAgenticKnowledge
Reasoning
92.5#2 of 16 · 2 tests
Coding
83.71 test · not ranked
Agentic
90.5#2 of 8 · 2 tests
Knowledge
99.11 test · not ranked

Model facts

Lab
OpenAI
Type
Closed API
Released
2026-09-03
Context
1.05M
API price / 1M tokens
$10 in · $50 out

GPT-6 Astra pricing and open alternatives

GPT-6 Astra is only available through OpenAI's API, so it cannot be self-hosted; it costs $10 per million input tokens and $50 per million output tokens.

Open-weight alternatives

The highest-ranked open models, which you can run on your own GPU.

GPT-6 Astra speed and pricing

Measured figures link to their source; estimates are labelled.

API output speed
74 tok/sArtificial Analysis median · Artificial Analysis
Time to first token
3.86 sOpenRouter model page provider stats
API price per 1M tokens
$10.00 in · $50.00 outlab list price · OpenAI API pricing docs
Compare speed and price across all models

Score by test

Rank is among every model with a published score on that test.

GPQA Diamond96.1%GPQA Diamond; AA independent run (Max)#1 of 19 · top scoreHumanity's Last Exam54.7%Humanity's Last Exam, no tools; AA independent run (Max)#5 of 23 · 9.7 behind Claude Opus 5.5Humanity's Last Exam (with tools)57.2%Humanity's Last Exam, with tools; max score at any effort#10 of 17 · 12.8 behind Kimi K3.1 ProTerminal-Bench 4.059.1%Terminal-Bench 4.0; AA independent run (Max)#4 of 16 · 11.5 behind Claude Sonnet 5.5BrowseComp91.5%BrowseComp; max score at any effort#2 of 8 · 0.6 behind Kimi K3.1 ProMMMU-Pro86.9%MMMU-Pro (not original MMMU); AA independent run (Max)#2 of 9 · 0.8 behind Claude Opus 5.5LMArena Text Elo1476LMArena Text, Overall; snapshot Oct 27, 2026; GPT-6 Astra (Max); +/-6#12 of 21 · 44 behind Gemini 4 Argon

Nearby in the composite ranking

Models with the closest composite index.

  1. 1Claude Opus 5.598.4
  2. 2Kimi K3.1 Pro94.4
  3. 3Claude Sonnet 5.591.9
  4. 4GPT-6 Astra91.4
  5. 5Claude Fable 5.191.2
  6. 6DeepSeek V4 Pro91.0
  7. 7Kimi K390.6

Head-to-head comparisons

Side-by-side pages built from the tests both models report.

Sources

ModelTestScoreSettingSource
GPT-6 AstraGPQA Diamond96.1GPQA Diamond; AA independent run (Max)Artificial Analysis
GPT-6 AstraHumanity's Last Exam54.7Humanity's Last Exam, no tools; AA independent run (Max)Artificial Analysis
GPT-6 AstraHumanity's Last Exam (with tools)57.2Humanity's Last Exam, with tools; max score at any effortOpenAI GPT-6 Astra launch post (Academic table)
GPT-6 AstraTerminal-Bench 4.059.1Terminal-Bench 4.0; AA independent run (Max)Artificial Analysis
GPT-6 AstraBrowseComp91.5BrowseComp; max score at any effortOpenAI GPT-6 Astra launch post (Professional table)
GPT-6 AstraMMMU-Pro86.9MMMU-Pro (not original MMMU); AA independent run (Max)Artificial Analysis
GPT-6 AstraLMArena Text Elo1476LMArena Text, Overall; snapshot Oct 27, 2026; GPT-6 Astra (Max); +/-6LMArena (arena.ai) Text leaderboard

Scores are published results from model cards, lab reports, and public leaderboards, each linked to its source. Settings such as thinking mode or test version differ between sources and are shown next to each score. OpenLLM Buddy catalog models are uncensored fine-tunes; their scores are those of the official base model. Data collected 2026-10-30. How the index works