Skip to content

Benchmark

Proofbench

Scores are percentages.

Models measured
83
Scores recorded
84
Sources
1
IQ weight
2.12

Score distribution

41.9%
Mean
36%
Median
31.7%
Std dev
5%
10th percentile
95%
90th percentile
0%
Min
0%100%
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 2.12 for chat models, measured over 83 models. The weight is how much it counts toward llminfo IQ. We measure it from the data, not from the raw scores.

Leaderboard

#ModelScore
1Claude Fable 5.1max· Epoch AI Benchmarking Hub data100%
2Claude Fable 5.2max· Epoch AI Benchmarking Hub data100%
3Claude Opus 5.5max· Epoch AI Benchmarking Hub data100%
4Claude Sonnet 5.5max· Epoch AI Benchmarking Hub data100%
5Claude Opus 5max· Epoch AI Benchmarking Hub data99%
6Gemini 4 Argon· Epoch AI Benchmarking Hub data99%
7GPT-6 Astra· Epoch AI Benchmarking Hub data99%
8GPT-6.1 Sol· Epoch AI Benchmarking Hub data99%
9Claude Fable 5max· Epoch AI Benchmarking Hub data95%
10Kimi K3.1 Pro· Epoch AI Benchmarking Hub data91%
11Claude Haiku 5.5max· Epoch AI Benchmarking Hub data89%
12Kimi K3· Epoch AI Benchmarking Hub data87%
13Muse Wonder 1.0· Epoch AI Benchmarking Hub data84%
14GPT-5.6 Solmax· Epoch AI Benchmarking Hub data83%
15GPT-6 Sol· Epoch AI Benchmarking Hub data83%
16Claude Sonnet 5max· Epoch AI Benchmarking Hub data77%
17Hy4 preview· Epoch AI Benchmarking Hub data75%
18GPT-5.6 Terraxhigh· Epoch AI Benchmarking Hub data74%
19MiMo V2.6 Pro· Epoch AI Benchmarking Hub data70%
20Claude Opus 4.8max· Epoch AI Benchmarking Hub data69%
21GPT-6 Luna· Epoch AI Benchmarking Hub data64%
22MiMo V2.6 Flash· Epoch AI Benchmarking Hub data63%
23GPT-5.6 Lunamax· Epoch AI Benchmarking Hub data60%
24Gemini 3.7 Flash· Epoch AI Benchmarking Hub data58%
25Muse Spark 1.3max· Epoch AI Benchmarking Hub data58%
26Qwen3.8 Max· Epoch AI Benchmarking Hub data58%
27DeepSeek V4 Flash 0731· Epoch AI Benchmarking Hub data56%
28GPT 5.4xhigh· Epoch AI Benchmarking Hub data56%
29Muse Spark 1.3· Epoch AI Benchmarking Hub data55%
30Claude Opus 4.7max· Epoch AI Benchmarking Hub data54%
31DeepSeek V4.1 Flash· Epoch AI Benchmarking Hub data54%
32Grok 4.6· Epoch AI Benchmarking Hub data51%
33Claude Opus 4.6max· Epoch AI Benchmarking Hub data50%
34DeepSeek V4 Pro 0813· Epoch AI Benchmarking Hub data50%
35GPT 5.5xhigh· Epoch AI Benchmarking Hub data50%
36GLM 5.3max· Epoch AI Benchmarking Hub data49%
37Gemini 3.8 Flash· Epoch AI Benchmarking Hub data48%
38Claude Sonnet 4.6max· Epoch AI Benchmarking Hub data45%
39Muse Spark 1.2· Epoch AI Benchmarking Hub data43%
40Step 5 Preview· Epoch AI Benchmarking Hub data42%
41Muse Spark 1.1· Epoch AI Benchmarking Hub data39%
42Claude Opus 4.5· Epoch AI Benchmarking Hub data36%
43Gemini 3.6 Flash· Epoch AI Benchmarking Hub data36%
44GLM 5.2max· Epoch AI Benchmarking Hub data35%
45Gemini 3.5 Flashhigh· Epoch AI Benchmarking Hub data31%
46Grok 4.5high· Epoch AI Benchmarking Hub data31%
47Gemini 3.1 Pro Preview· Epoch AI Benchmarking Hub data26%
48Grok 4.7· Epoch AI Benchmarking Hub data26%
49Qwen3.7 Max· Epoch AI Benchmarking Hub data26%
50GLM 5.1· Epoch AI Benchmarking Hub data22.2%

Showing 1–50 of 84 results.

Data credits: Epoch AI Benchmarking Hub, CC BY 4.0 (licence)