OpenAI, released Sep 3, 2026

GPT-6 Astraprice, context, benchmarks and release details

100% confidence 100 percent, Full confidence, 7 of 7 expected sources in
77.5
SI Score
#3 of 142 ranked models
Input, per 1M tokens
$10.00OpenAI pricingOfficial Standard short-context rate; excludes Batch/Flex/cache discounts Retrieved Oct 9, 2026 · factual citation
Open source ↗
Output, per 1M tokens
$50.00OpenAI pricingOfficial Standard short-context rate; excludes Batch/Flex/cache discounts Retrieved Oct 9, 2026 · factual citation
Open source ↗
Context window
1.1Mmodels.devPublished source fact Retrieved Oct 9, 2026 · MIT
Open source ↗
Max output
128Kmodels.devPublished source fact Retrieved Oct 9, 2026 · MIT
Open source ↗
Released
Sep 3, 2026OpenAI API changelogPublished source fact Retrieved Oct 9, 2026 · factual citation
Open source ↗

How the score breaks down

Coding (weight 40 percent) 64.8
Math (weight 15 percent) 92.9
Preference (weight 15 percent) 76.9
Reasoning (weight 30 percent) 87.1

Weights: reasoning 30%, math 15%, coding 40%, preference 15%. Results use fixed 0–100 scales before averaging, and thin evidence is pulled toward 50. Method si-v3-retained-evidence-2, computed Oct 9, 2026, 05:13 UTC.

Around it on the leaderboard

  1. 1 Claude Fable 5.1 80.2
  2. 2 Claude Opus 5.5 78.4
  3. 3 GPT-6 Astra 77.5
  4. 4 Claude Fable 5 76.8
  5. 5 Claude Opus 5 74.9

Full leaderboard

Benchmark results

32 benchmarks, 63 results

Each row shows the best published result. Where a model was tested at several settings, such as reasoning effort, open the row to see each one. Hover or tap a value for its source.

ARC-AGI-1 (public eval)reasoning 99%ARC PrizeARC Prize steward-published result; exact edition/subset; publication time is HTTP Last-Modified of the aggregate export, not the evaluation date. Evaluation conducted with 'xhigh' reasoning effort. [variant] openai-gpt-6-astra-xhighPublished Oct 6, 2026 Retrieved Oct 9, 2026 · factual citation
Open source ↗ 99.0 5 settings
  • low effort 98.3%ARC PrizeARC Prize steward-published result; exact edition/subset; publication time is HTTP Last-Modified of the aggregate export, not the evaluation date. Evaluation conducted with 'low' reasoning effort. [variant] openai-gpt-6-astra-lowPublished Oct 6, 2026 Retrieved Oct 9, 2026 · factual citation
    Open source ↗
  • medium effort 98.8%ARC PrizeARC Prize steward-published result; exact edition/subset; publication time is HTTP Last-Modified of the aggregate export, not the evaluation date. Evaluation conducted with 'medium' reasoning effort. [variant] openai-gpt-6-astra-mediumPublished Oct 6, 2026 Retrieved Oct 9, 2026 · factual citation
    Open source ↗
  • high effort 98.5%ARC PrizeARC Prize steward-published result; exact edition/subset; publication time is HTTP Last-Modified of the aggregate export, not the evaluation date. Evaluation conducted with 'high' reasoning effort. [variant] openai-gpt-6-astra-highPublished Oct 6, 2026 Retrieved Oct 9, 2026 · factual citation
    Open source ↗
  • xhigh effort 99%ARC PrizeARC Prize steward-published result; exact edition/subset; publication time is HTTP Last-Modified of the aggregate export, not the evaluation date. Evaluation conducted with 'xhigh' reasoning effort. [variant] openai-gpt-6-astra-xhighPublished Oct 6, 2026 Retrieved Oct 9, 2026 · factual citation
    Open source ↗
  • max effort 97.8%ARC PrizeARC Prize steward-published result; exact edition/subset; publication time is HTTP Last-Modified of the aggregate export, not the evaluation date. Evaluation conducted with 'max' reasoning effort. [variant] openai-gpt-6-astra-maxPublished Oct 6, 2026 Retrieved Oct 9, 2026 · factual citation
    Open source ↗
About ARC-AGI-1 (public eval)
ARC-AGI-1 (semi-private)reasoning 98.5%ARC PrizeARC Prize steward-published result; exact edition/subset; publication time is HTTP Last-Modified of the aggregate export, not the evaluation date. Evaluation conducted with 'high' reasoning effort. [variant] openai-gpt-6-astra-highPublished Oct 6, 2026 Retrieved Oct 9, 2026 · factual citation
Open source ↗ 98.5 5 settings
  • low effort 96.5%ARC PrizeARC Prize steward-published result; exact edition/subset; publication time is HTTP Last-Modified of the aggregate export, not the evaluation date. Evaluation conducted with 'low' reasoning effort. [variant] openai-gpt-6-astra-lowPublished Oct 6, 2026 Retrieved Oct 9, 2026 · factual citation
    Open source ↗
  • medium effort 97.5%ARC PrizeARC Prize steward-published result; exact edition/subset; publication time is HTTP Last-Modified of the aggregate export, not the evaluation date. Evaluation conducted with 'medium' reasoning effort. [variant] openai-gpt-6-astra-mediumPublished Oct 6, 2026 Retrieved Oct 9, 2026 · factual citation
    Open source ↗
  • high effort 98.5%ARC PrizeARC Prize steward-published result; exact edition/subset; publication time is HTTP Last-Modified of the aggregate export, not the evaluation date. Evaluation conducted with 'high' reasoning effort. [variant] openai-gpt-6-astra-highPublished Oct 6, 2026 Retrieved Oct 9, 2026 · factual citation
    Open source ↗
  • xhigh effort 98.5%ARC PrizeARC Prize steward-published result; exact edition/subset; publication time is HTTP Last-Modified of the aggregate export, not the evaluation date. Evaluation conducted with 'xhigh' reasoning effort. [variant] openai-gpt-6-astra-xhighPublished Oct 6, 2026 Retrieved Oct 9, 2026 · factual citation
    Open source ↗
  • max effort 97.5%ARC PrizeARC Prize steward-published result; exact edition/subset; publication time is HTTP Last-Modified of the aggregate export, not the evaluation date. Evaluation conducted with 'max' reasoning effort. [variant] openai-gpt-6-astra-maxPublished Oct 6, 2026 Retrieved Oct 9, 2026 · factual citation
    Open source ↗
About ARC-AGI-1 (semi-private)
ARC-AGI-2 (public eval)reasoning 97.9%ARC PrizeARC Prize steward-published result; exact edition/subset; publication time is HTTP Last-Modified of the aggregate export, not the evaluation date. Evaluation conducted with 'high' reasoning effort. [variant] openai-gpt-6-astra-highPublished Oct 6, 2026 Retrieved Oct 9, 2026 · factual citation
Open source ↗ 97.9 5 settings
  • low effort 94.2%ARC PrizeARC Prize steward-published result; exact edition/subset; publication time is HTTP Last-Modified of the aggregate export, not the evaluation date. Evaluation conducted with 'low' reasoning effort. [variant] openai-gpt-6-astra-lowPublished Oct 6, 2026 Retrieved Oct 9, 2026 · factual citation
    Open source ↗
  • medium effort 96.7%ARC PrizeARC Prize steward-published result; exact edition/subset; publication time is HTTP Last-Modified of the aggregate export, not the evaluation date. Evaluation conducted with 'medium' reasoning effort. [variant] openai-gpt-6-astra-mediumPublished Oct 6, 2026 Retrieved Oct 9, 2026 · factual citation
    Open source ↗
  • high effort 97.9%ARC PrizeARC Prize steward-published result; exact edition/subset; publication time is HTTP Last-Modified of the aggregate export, not the evaluation date. Evaluation conducted with 'high' reasoning effort. [variant] openai-gpt-6-astra-highPublished Oct 6, 2026 Retrieved Oct 9, 2026 · factual citation
    Open source ↗
  • xhigh effort 97.9%ARC PrizeARC Prize steward-published result; exact edition/subset; publication time is HTTP Last-Modified of the aggregate export, not the evaluation date. Evaluation conducted with 'xhigh' reasoning effort. [variant] openai-gpt-6-astra-xhighPublished Oct 6, 2026 Retrieved Oct 9, 2026 · factual citation
    Open source ↗
  • max effort 97.5%ARC PrizeARC Prize steward-published result; exact edition/subset; publication time is HTTP Last-Modified of the aggregate export, not the evaluation date. Evaluation conducted with 'max' reasoning effort. [variant] openai-gpt-6-astra-maxPublished Oct 6, 2026 Retrieved Oct 9, 2026 · factual citation
    Open source ↗
About ARC-AGI-2 (public eval)
ARC-AGI-2 (semi-private)reasoning 95%ARC PrizeARC Prize steward-published result; exact edition/subset; publication time is HTTP Last-Modified of the aggregate export, not the evaluation date. Evaluation conducted with 'max' reasoning effort. [variant] openai-gpt-6-astra-maxPublished Oct 6, 2026 Retrieved Oct 9, 2026 · factual citation
Open source ↗ 95.0 5 settings
  • low effort 85.4%ARC PrizeARC Prize steward-published result; exact edition/subset; publication time is HTTP Last-Modified of the aggregate export, not the evaluation date. Evaluation conducted with 'low' reasoning effort. [variant] openai-gpt-6-astra-lowPublished Oct 6, 2026 Retrieved Oct 9, 2026 · factual citation
    Open source ↗
  • medium effort 92.1%ARC PrizeARC Prize steward-published result; exact edition/subset; publication time is HTTP Last-Modified of the aggregate export, not the evaluation date. Evaluation conducted with 'medium' reasoning effort. [variant] openai-gpt-6-astra-mediumPublished Oct 6, 2026 Retrieved Oct 9, 2026 · factual citation
    Open source ↗
  • high effort 92.1%ARC PrizeARC Prize steward-published result; exact edition/subset; publication time is HTTP Last-Modified of the aggregate export, not the evaluation date. Evaluation conducted with 'high' reasoning effort. [variant] openai-gpt-6-astra-highPublished Oct 6, 2026 Retrieved Oct 9, 2026 · factual citation
    Open source ↗
  • xhigh effort 93.3%ARC PrizeARC Prize steward-published result; exact edition/subset; publication time is HTTP Last-Modified of the aggregate export, not the evaluation date. Evaluation conducted with 'xhigh' reasoning effort. [variant] openai-gpt-6-astra-xhighPublished Oct 6, 2026 Retrieved Oct 9, 2026 · factual citation
    Open source ↗
  • max effort 95%ARC PrizeARC Prize steward-published result; exact edition/subset; publication time is HTTP Last-Modified of the aggregate export, not the evaluation date. Evaluation conducted with 'max' reasoning effort. [variant] openai-gpt-6-astra-maxPublished Oct 6, 2026 Retrieved Oct 9, 2026 · factual citation
    Open source ↗
About ARC-AGI-2 (semi-private)
ARC-AGI-3reasoning 99.9%Official model cards via models.devLab-reported; metric RHAE; transcribed by MIT models.dev catalog; not independently evaluated [variant] Responses APIPublished Sep 3, 2026 Retrieved Oct 9, 2026 · factual citation; MIT transcription
Open source ↗ 99.9
ARC-AGI-3 (semi-private)reasoning 62.7%ARC PrizeARC Prize steward-published result; exact edition/subset; publication time is HTTP Last-Modified of the aggregate export, not the evaluation date. Evaluation conducted with 'max' reasoning effort. [variant] openai-gpt-6-astra-maxPublished Oct 6, 2026 Retrieved Oct 9, 2026 · factual citation
Open source ↗ 62.7 5 settings
  • low effort 17.5%ARC PrizeARC Prize steward-published result; exact edition/subset; publication time is HTTP Last-Modified of the aggregate export, not the evaluation date. Evaluation conducted with 'low' reasoning effort. [variant] openai-gpt-6-astra-lowPublished Oct 6, 2026 Retrieved Oct 9, 2026 · factual citation
    Open source ↗
  • medium effort 38.6%ARC PrizeARC Prize steward-published result; exact edition/subset; publication time is HTTP Last-Modified of the aggregate export, not the evaluation date. Evaluation conducted with 'medium' reasoning effort. [variant] openai-gpt-6-astra-mediumPublished Oct 6, 2026 Retrieved Oct 9, 2026 · factual citation
    Open source ↗
  • high effort 54.8%ARC PrizeARC Prize steward-published result; exact edition/subset; publication time is HTTP Last-Modified of the aggregate export, not the evaluation date. Evaluation conducted with 'high' reasoning effort. [variant] openai-gpt-6-astra-highPublished Oct 6, 2026 Retrieved Oct 9, 2026 · factual citation
    Open source ↗
  • xhigh effort 59.3%ARC PrizeARC Prize steward-published result; exact edition/subset; publication time is HTTP Last-Modified of the aggregate export, not the evaluation date. Evaluation conducted with 'xhigh' reasoning effort. [variant] openai-gpt-6-astra-xhighPublished Oct 6, 2026 Retrieved Oct 9, 2026 · factual citation
    Open source ↗
  • max effort 62.7%ARC PrizeARC Prize steward-published result; exact edition/subset; publication time is HTTP Last-Modified of the aggregate export, not the evaluation date. Evaluation conducted with 'max' reasoning effort. [variant] openai-gpt-6-astra-maxPublished Oct 6, 2026 Retrieved Oct 9, 2026 · factual citation
    Open source ↗
About ARC-AGI-3 (semi-private)
GPQA Diamondreasoning 96%Official model cards via models.devLab-reported; metric accuracy; transcribed by MIT models.dev catalog; not independently evaluated [variant] Published Sep 3, 2026 Retrieved Oct 9, 2026 · factual citation; MIT transcription
Open source ↗ 96.0 2 settings
  • max effort 95.8%Epoch AI BenchmarkingEpoch-owned evaluation mean score; scores only, no benchmark questions [variant] maxPublished Aug 30, 2026 Retrieved Oct 9, 2026 · CC-BY
    Open source ↗
  • Setting 2 96%Official model cards via models.devLab-reported; metric accuracy; transcribed by MIT models.dev catalog; not independently evaluated [variant] Published Sep 3, 2026 Retrieved Oct 9, 2026 · factual citation; MIT transcription
    Open source ↗
About GPQA Diamond
Humanity's Last Exam (Scale AI)reasoning 54.8%Humanity’s Last ExamPotential contamination warning: This model was evaluated after the public release of HLE, allowing model builder access to the prompts and solutions. [variant] Published Sep 9, 2026 Retrieved Oct 9, 2026 · factual citation
Open source ↗ 54.8
Humanity's Last Exam (with tools)reasoning 57.2%Official model cards via models.devLab-reported; metric accuracy; transcribed by MIT models.dev catalog; not independently evaluated [variant] with toolsPublished Sep 3, 2026 Retrieved Oct 9, 2026 · factual citation; MIT transcription
Open source ↗ 57.2
LiveBench Reasoning: connectionsreasoning 100%LiveBenchLiveBench subtask result; dataset edition 2026_06_25; publication time is HTTP Last-Modified of the score CSV, not the dataset edition or individual evaluation date [variant] 2026_06_25Published Oct 7, 2026 Retrieved Oct 9, 2026 · factual citation; Apache-2.0 code
Open source ↗ 100.0
LiveBench Reasoning: consecutive eventsreasoning 90.6%LiveBenchLiveBench subtask result; dataset edition 2026_06_25; publication time is HTTP Last-Modified of the score CSV, not the dataset edition or individual evaluation date [variant] 2026_06_25Published Oct 7, 2026 Retrieved Oct 9, 2026 · factual citation; Apache-2.0 code
Open source ↗ 90.6
LiveBench Reasoning: logic with navigationreasoning 88%LiveBenchLiveBench subtask result; dataset edition 2026_06_25; publication time is HTTP Last-Modified of the score CSV, not the dataset edition or individual evaluation date [variant] 2026_06_25Published Oct 7, 2026 Retrieved Oct 9, 2026 · factual citation; Apache-2.0 code
Open source ↗ 88.0
LiveBench Reasoning: spatialreasoning 98%LiveBenchLiveBench subtask result; dataset edition 2026_06_25; publication time is HTTP Last-Modified of the score CSV, not the dataset edition or individual evaluation date [variant] 2026_06_25Published Oct 7, 2026 Retrieved Oct 9, 2026 · factual citation; Apache-2.0 code
Open source ↗ 98.0
LiveBench Reasoning: theory of mindreasoning 84.6%LiveBenchLiveBench subtask result; dataset edition 2026_06_25; publication time is HTTP Last-Modified of the score CSV, not the dataset edition or individual evaluation date [variant] 2026_06_25Published Oct 7, 2026 Retrieved Oct 9, 2026 · factual citation; Apache-2.0 code
Open source ↗ 84.6
LiveBench Reasoning: zebra puzzlesreasoning 100%LiveBenchLiveBench subtask result; dataset edition 2026_06_25; publication time is HTTP Last-Modified of the score CSV, not the dataset edition or individual evaluation date [variant] 2026_06_25Published Oct 7, 2026 Retrieved Oct 9, 2026 · factual citation; Apache-2.0 code
Open source ↗ 100.0
FrontierMath Tier 4 (v2)math 97.6%Epoch AI BenchmarkingEpoch-owned evaluation mean score; scores only, no benchmark questions [variant] highPublished Aug 30, 2026 Retrieved Oct 9, 2026 · CC-BY
Open source ↗ 97.6 6 settings
  • no reasoning 82.9%Epoch AI BenchmarkingEpoch-owned evaluation mean score; scores only, no benchmark questions [variant] nonePublished Aug 30, 2026 Retrieved Oct 9, 2026 · CC-BY
    Open source ↗
  • low effort 87.8%Epoch AI BenchmarkingEpoch-owned evaluation mean score; scores only, no benchmark questions [variant] lowPublished Aug 30, 2026 Retrieved Oct 9, 2026 · CC-BY
    Open source ↗
  • medium effort 97.6%Epoch AI BenchmarkingEpoch-owned evaluation mean score; scores only, no benchmark questions [variant] mediumPublished Aug 30, 2026 Retrieved Oct 9, 2026 · CC-BY
    Open source ↗
  • high effort 97.6%Epoch AI BenchmarkingEpoch-owned evaluation mean score; scores only, no benchmark questions [variant] highPublished Aug 30, 2026 Retrieved Oct 9, 2026 · CC-BY
    Open source ↗
  • xhigh effort 97.6%Epoch AI BenchmarkingEpoch-owned evaluation mean score; scores only, no benchmark questions [variant] xhighPublished Aug 30, 2026 Retrieved Oct 9, 2026 · CC-BY
    Open source ↗
  • max effort 97.6%Epoch AI BenchmarkingEpoch-owned evaluation mean score; scores only, no benchmark questions [variant] maxPublished Aug 30, 2026 Retrieved Oct 9, 2026 · CC-BY
    Open source ↗
About FrontierMath Tier 4 (v2)
FrontierMath Tier 4 (v2)math 97.6%Official model cards via models.devLab-reported; metric accuracy; transcribed by MIT models.dev catalog; not independently evaluated [variant] Tier 4; v2Published Sep 3, 2026 Retrieved Oct 9, 2026 · factual citation; MIT transcription
Open source ↗ 97.6
FrontierMath Tiers 1–3 (v2)math 93.7%Epoch AI BenchmarkingEpoch-owned evaluation mean score; scores only, no benchmark questions [variant] maxPublished Aug 30, 2026 Retrieved Oct 9, 2026 · CC-BY
Open source ↗ 93.7
LiveBench Math: AMPS Hardmath 98%LiveBenchLiveBench subtask result; dataset edition 2026_06_25; publication time is HTTP Last-Modified of the score CSV, not the dataset edition or individual evaluation date [variant] 2026_06_25Published Oct 7, 2026 Retrieved Oct 9, 2026 · factual citation; Apache-2.0 code
Open source ↗ 98.0
LiveBench Math: competition mathmath 97.1%LiveBenchLiveBench subtask result; dataset edition 2026_06_25; publication time is HTTP Last-Modified of the score CSV, not the dataset edition or individual evaluation date [variant] 2026_06_25Published Oct 7, 2026 Retrieved Oct 9, 2026 · factual citation; Apache-2.0 code
Open source ↗ 97.1
LiveBench Math: integralsmath 100%LiveBenchLiveBench subtask result; dataset edition 2026_06_25; publication time is HTTP Last-Modified of the score CSV, not the dataset edition or individual evaluation date [variant] 2026_06_25Published Oct 7, 2026 Retrieved Oct 9, 2026 · factual citation; Apache-2.0 code
Open source ↗ 100.0
LiveBench Math: olympiadmath 92.2%LiveBenchLiveBench subtask result; dataset edition 2026_06_25; publication time is HTTP Last-Modified of the score CSV, not the dataset edition or individual evaluation date [variant] 2026_06_25Published Oct 7, 2026 Retrieved Oct 9, 2026 · factual citation; Apache-2.0 code
Open source ↗ 92.2
LiveBench Math: simplifymath 70.5%LiveBenchLiveBench subtask result; dataset edition 2026_06_25; publication time is HTTP Last-Modified of the score CSV, not the dataset edition or individual evaluation date [variant] 2026_06_25Published Oct 7, 2026 Retrieved Oct 9, 2026 · factual citation; Apache-2.0 code
Open source ↗ 70.5
OTIS Mock AIME 2024–2025math 100%Epoch AI BenchmarkingEpoch-owned evaluation mean score; scores only, no benchmark questions [variant] maxPublished Aug 30, 2026 Retrieved Oct 9, 2026 · CC-BY
Open source ↗ 100.0
LiveBench Coding: code completioncoding 80.4%LiveBenchLiveBench subtask result; dataset edition 2026_06_25; publication time is HTTP Last-Modified of the score CSV, not the dataset edition or individual evaluation date [variant] 2026_06_25Published Oct 7, 2026 Retrieved Oct 9, 2026 · factual citation; Apache-2.0 code
Open source ↗ 80.4
LiveBench Coding: code generationcoding 80.3%LiveBenchLiveBench subtask result; dataset edition 2026_06_25; publication time is HTTP Last-Modified of the score CSV, not the dataset edition or individual evaluation date [variant] 2026_06_25Published Oct 7, 2026 Retrieved Oct 9, 2026 · factual citation; Apache-2.0 code
Open source ↗ 80.3
LiveBench Coding: JavaScriptcoding 63.6%LiveBenchLiveBench subtask result; dataset edition 2026_06_25; publication time is HTTP Last-Modified of the score CSV, not the dataset edition or individual evaluation date [variant] 2026_06_25Published Oct 7, 2026 Retrieved Oct 9, 2026 · factual citation; Apache-2.0 code
Open source ↗ 63.6
LiveBench Coding: Pythoncoding 65%LiveBenchLiveBench subtask result; dataset edition 2026_06_25; publication time is HTTP Last-Modified of the score CSV, not the dataset edition or individual evaluation date [variant] 2026_06_25Published Oct 7, 2026 Retrieved Oct 9, 2026 · factual citation; Apache-2.0 code
Open source ↗ 65.0
LiveBench Coding: TypeScriptcoding 43.3%LiveBenchLiveBench subtask result; dataset edition 2026_06_25; publication time is HTTP Last-Modified of the score CSV, not the dataset edition or individual evaluation date [variant] 2026_06_25Published Oct 7, 2026 Retrieved Oct 9, 2026 · factual citation; Apache-2.0 code
Open source ↗ 43.3
Terminal-Bench 0.1coding 64.6%Official model cards via models.devLab-reported; metric success rate; transcribed by MIT models.dev catalog; not independently evaluated [variant] 0.1Published Sep 3, 2026 Retrieved Oct 9, 2026 · factual citation; MIT transcription
Open source ↗ 64.6
Terminal-Bench 4.0coding 58.2%Terminal-BenchTerminal-Bench 4.0; published harness submission, 95% CI retained at source [variant] Codex; maxPublished Sep 10, 2026 Retrieved Oct 9, 2026 · Apache-2.0; factual citation
Open source ↗ 58.2 6 settings
  • Setting 1 57.9%Official model cards via models.devLab-reported; metric success rate; transcribed by MIT models.dev catalog; not independently evaluated [variant] 4.0Published Sep 3, 2026 Retrieved Oct 9, 2026 · factual citation; MIT transcription
    Open source ↗
  • Setting 2 57.9%Terminal-BenchTerminal-Bench 4.0; published harness submission, 95% CI retained at source [variant] Codex; highPublished Sep 10, 2026 Retrieved Oct 9, 2026 · Apache-2.0; factual citation
    Open source ↗
  • Setting 3 50.6%Terminal-BenchTerminal-Bench 4.0; published harness submission, 95% CI retained at source [variant] Codex; lowPublished Sep 10, 2026 Retrieved Oct 9, 2026 · Apache-2.0; factual citation
    Open source ↗
  • Setting 4 58.2%Terminal-BenchTerminal-Bench 4.0; published harness submission, 95% CI retained at source [variant] Codex; maxPublished Sep 10, 2026 Retrieved Oct 9, 2026 · Apache-2.0; factual citation
    Open source ↗
  • Setting 5 54.2%Terminal-BenchTerminal-Bench 4.0; published harness submission, 95% CI retained at source [variant] Codex; mediumPublished Sep 10, 2026 Retrieved Oct 9, 2026 · Apache-2.0; factual citation
    Open source ↗
  • Setting 6 57.9%Terminal-BenchTerminal-Bench 4.0; published harness submission, 95% CI retained at source [variant] Codex; xhighPublished Sep 10, 2026 Retrieved Oct 9, 2026 · Apache-2.0; factual citation
    Open source ↗
About Terminal-Bench 4.0
LMArena Textpreference 1440.5 eloLMArena / ArenaPublished source fact [variant] text / overallPublished Oct 8, 2026 Retrieved Oct 9, 2026 · CC-BY-4.0
Open source ↗ 76.9

Normalization uses a fixed 0–100 scale for each unit, independent of other models. Compare evaluation conditions before reading a small gap as decisive. “Lab-reported” marks the provider's own published figure.

Details and sources

Open weights
Nomodels.devPublished source fact Retrieved Oct 9, 2026 · MIT
Open source ↗
License
not yet reported
Input modalities
text, image, pdfmodels.devPublished source fact Retrieved Oct 9, 2026 · MIT
Open source ↗
First seen by SuperIndex
Oct 8, 2026
Coverage
100% of expected source weight

Reported (7)

  • ARC PrizeOct 8, 2026
  • Epoch AI BenchmarkingOct 8, 2026
  • Humanity’s Last ExamOct 8, 2026
  • Official model cards via models.devOct 8, 2026
  • LiveBenchOct 8, 2026
  • LMArena / ArenaOct 8, 2026
  • Terminal-BenchOct 8, 2026

Awaiting (0)

Every expected source has reported for this model.

Confidence rises as pending sources publish. Some sources never cover some models, so confidence reaches 100% at 80% of expected weight.

Compare GPT-6 Astra

All comparisons
vs Claude Fable 5.1 Anthropic · SI 80.2 vs Claude Opus 5.5 Anthropic · SI 78.4 vs Claude Fable 5 Anthropic · SI 76.8 vs Claude Opus 5 Anthropic · SI 74.9

What changed

All releases · RSS feed

Alerts on this device

What to be alerted about
RSS feed