Noometry

Reasoning benchmark

EBR-Bench leaderboard

As of October 2026, GPT-6 Astra has the highest published EBR-Bench score on Noometry at 76.2%, out of 24 models with results.

Last verified

About EBR-Bench

A description with primary sources is being prepared for this benchmark.

Category
Reasoning
Introduced
2026
Unit
Percent (random guessing ≈ 0%)
Official site
epoch.ai

Top 15 models

  1. GPT-6 Astra 76.2%
  2. Claude Opus 5.5 71.4%
  3. Claude Fable 5.1 57.1%
  4. GPT-6.1 Sol 54.3%
  5. GPT-6 Sol 53.3%
  6. Claude Opus 5 45.7%
  7. GPT-5.6 Sol 44.8%
  8. Claude Fable 5 39.5%
  9. GPT-5.5 34.3%
  10. Grok 4.6 30.5%
  11. Claude Opus 4.8 28.6%
  12. GPT-5.4 25.4%
  13. GPT-5.2 23%
  14. Claude Opus 4.7 19%
  15. Claude Opus 4.5 14.3%

Sponsored placements are available on pages like this one. Advertise on Noometry

All results

#ModelProviderScoreSettingSourceDate
1GPT-6 Astra OpenAI76.2%maxEpoch AI2026-08-30
2Claude Opus 5.5 Anthropic71.4%maxEpoch AI2026-09-22
3Claude Fable 5.1 Anthropic57.1%maxEpoch AI2026-09-08
4GPT-6.1 Sol OpenAI54.3%maxEpoch AI2026-09-29
5GPT-6 Sol OpenAI53.3%maxEpoch AI2026-09-22
6Claude Opus 5 Anthropic45.7%maxEpoch AI2026-09-05
7GPT-5.6 Sol OpenAI44.8%maxEpoch AI2026-09-04
8Claude Fable 5 Anthropic39.5%maxEpoch AI2026-07-17
9GPT-5.5 OpenAI34.3%xhighEpoch AI2026-07-27
10Grok 4.6 xAI30.5%xhighEpoch AI2026-08-18
11Claude Opus 4.8 Anthropic28.6%maxEpoch AI2026-08-07
12GPT-5.4 OpenAI25.4%xhighEpoch AI2026-06-25
13GPT-5.2 OpenAI23%xhighEpoch AI2026-06-26
14Claude Opus 4.7 Anthropic19%maxEpoch AI2026-06-30
15Claude Opus 4.5 Anthropic14.3%128KEpoch AI2026-06-25
16Gemini 3.1 Pro Preview Google14.3%Epoch AI2026-06-25
17Claude Opus 4.6 Anthropic12.7%maxEpoch AI2026-06-29
18GPT-5 OpenAI12.7%highEpoch AI2026-06-29
19GLM-5.2 Z.ai (Zhipu)9.5%maxEpoch AI2026-06-29
20Qwen3.7 Max Alibaba (Qwen)9.5%Epoch AI2026-06-25
21Claude Opus 4.1 Anthropic7.9%Epoch AI2026-06-25
22Gemini 3.5 Flash Google4.8%highEpoch AI2026-06-25
23Claude Sonnet 4.5 Anthropic2.4%Epoch AI2026-06-25
24Kimi K2.6 Moonshot AI2.4%Epoch AI2026-06-25

Compare the leaders

Other reasoning benchmarks

Frequently asked questions

Which model has the highest EBR-Bench score?

As of October 2026, GPT-6 Astra has the highest published EBR-Bench score on Noometry at 76.2%, out of 24 models with results.

What is the best open-weight model on EBR-Bench?

GLM-5.2 has the highest EBR-Bench accuracy among open-weight models at 9.5%, ranking 19 of 24 overall.