Skip to content

Benchmark

FrontierMath Tier 4 (v2)

Math benchmarks measure how well models solve maths problems. Scores are percentages. Higher is better.

Models measured
74
Scores recorded
104
Sources
3
IQ weight
2.74

Score distribution

14.5%
Mean
0.561%
Median
29.1%
Std dev
0.049%
10th percentile
72.5%
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 2.74 for chat models, measured over 74 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 5.2max· Epoch AI Benchmarking Hub data100%
2GPT-6.1 Sol· LLM Stats data100%
3GPT-6.1 Solmax· Epoch AI Benchmarking Hub data100%
4GPT-6.2 Astramax· Epoch AI Benchmarking Hub data100%
5GPT-6 Astra· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim97.6%
6GPT-6 Astrahigh· Epoch AI Benchmarking Hub data97.6%
7GPT-6 Astramax· Epoch AI Benchmarking Hub data97.6%
8GPT-6 Astrav2 · GPT-6 Astra launch snapshot · 2026-09-01· Reported by OpenAI· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.97.6%
9GPT-6 Astraxhigh· Epoch AI Benchmarking Hub data97.6%
10Gemini 4 Neonxhigh· Epoch AI Benchmarking Hub data97.6%
11GPT-6 Astramedium· Epoch AI Benchmarking Hub data97.6%
12Gemini 4 Carbonhigh· Epoch AI Benchmarking Hub data95.1%
13Claude Opus 5.5max· Epoch AI Benchmarking Hub data95%
14GPT-6.1 Terramax· Epoch AI Benchmarking Hub data92.7%
15Claude Fable 5max· Epoch AI Benchmarking Hub data90.2%
16Claude Fable 5v2 · GPT-6 Astra launch snapshot · 2026-09-01· Reported by OpenAI· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.90.2%
17GPT-6 Solmax· Epoch AI Benchmarking Hub data90%
18Claude Fable 5.1max· Epoch AI Benchmarking Hub data87.8%
19Gemini 4 Hydrogenxhigh· Epoch AI Benchmarking Hub data87.8%
20GPT-6 Astralow· Epoch AI Benchmarking Hub data87.8%
21Claude Fable 5v2 · GPT-5.6 launch snapshot · 2026-07-09· Reported by OpenAI· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.87.8%
22Claude Fable 5.1v2 · GPT-6 Astra launch snapshot · 2026-09-01· Reported by OpenAI· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.87.8%
23GPT-5.6 Sol· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim83%
24GPT-5.6 Solv2 · GPT-5.6 launch snapshot · 2026-07-09· Reported by OpenAI· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.83%
25GPT-5.6 Solv2 · GPT-6 Astra launch snapshot · 2026-09-01· Reported by OpenAI· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.83%
26GPT-5.6 Solmax· Epoch AI Benchmarking Hub data82.9%
27GPT-6 Astranone· Epoch AI Benchmarking Hub data82.9%
28Claude Sonnet 5.5max· Epoch AI Benchmarking Hub data80.5%
29GPT-5.6 Solpro / max· Epoch AI Benchmarking Hub data80.5%
30Gemini 4 Argonhigh· Epoch AI Benchmarking Hub data78%
31GPT-5.5 Proxhigh· Epoch AI Benchmarking Hub data78%
32Claude Opus 5v2 · GPT-6 Astra launch snapshot · 2026-09-01· Reported by OpenAI· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.73.2%
33Claude Opus 5max· Epoch AI Benchmarking Hub data73.2%
34GPT 5.5v2 · GPT-5.6 launch snapshot · 2026-07-09· Reported by OpenAI· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.72.5%
35GPT 5.5xhigh· Epoch AI Benchmarking Hub data72.5%
36GPT-5.6 Terramax· Epoch AI Benchmarking Hub data70.7%
37Muse Wonder 1.0max· Epoch AI Benchmarking Hub data70.7%
38GPT-5.6 Terra· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim68.3%
39GPT-5.6 Terrav2 · GPT-5.6 launch snapshot · 2026-07-09· Reported by OpenAI· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.68.3%
40GPT-5.6 Lunamax· Epoch AI Benchmarking Hub data61%
41GPT-5.4 Proxhigh· Epoch AI Benchmarking Hub data58.5%
42GPT-5.6 Luna· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim58.5%
43GPT-5.6 Lunav2 · GPT-5.6 launch snapshot · 2026-07-09· Reported by OpenAI· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.58.5%
44Claude Opus 4.8v2 · GPT-5.6 launch snapshot · 2026-07-09· Reported by OpenAI· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.56.1%
45Claude Opus 4.8max· Epoch AI Benchmarking Hub data56.1%
46GPT-6 Lunamax· Epoch AI Benchmarking Hub data56.1%
47GPT 5.4xhigh· Epoch AI Benchmarking Hub data49%
48Claude Haiku 5.5max· Epoch AI Benchmarking Hub data46.3%
49DeepSeek V4.1 Flash 1017max· Epoch AI Benchmarking Hub data46.3%
50Muse Spark 1.3max· Epoch AI Benchmarking Hub data46.3%

Showing 1–50 of 104 results.

Data credits: BenchBoard by SYuan03 (github.com/SYuan03/benchboard). MIT covers its code and data organization; no results licence is inferred. Numeric relays retain the original stating-source links. · Epoch AI Benchmarking Hub, CC BY 4.0 (licence) · Data by LLM Stats