Skip to content

Benchmark

FrontierMath Tier 1-3 (v2)

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

Models measured
11
Scores recorded
13
Sources
1
IQ weight
0.75

Score distribution

65.2%
Mean
59.6%
Median
18.3%
Std dev
40.7%
10th percentile
87%
90th percentile
40.3%
Min
40.3%89%
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 0.75 for chat models, measured over 11 models. The weight is how much it counts toward llminfo IQ. We measure it from the data, not from the raw scores.

Leaderboard

#ModelScore
1GPT-5.6 Solv2· Reported by OpenAI· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.89%
2Claude Fable 5v2· Reported by OpenAI· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.87%
3GPT 5.5v2· Reported by OpenAI· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.85.3%
4GPT-5.6 Terrav2· Reported by OpenAI· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.84.9%
5Claude Opus 4.8v2· Reported by OpenAI· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.80%
6GPT-5.6 Lunav2· Reported by OpenAI· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.78.6%
7Gemini 3.1 Pro Previewv2· Reported by OpenAI· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.59.6%
8GPT-5.5 Prov2 · xhigh· Reported by OpenAI· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.52.4%
9GPT 5.5v2 · xhigh· Reported by OpenAI· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.51.7%
10GPT-5.4 ProOpenAI GPT-5.4 launch table· Reported by OpenAI· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.50%
11GPT 5.4OpenAI GPT-5.4 launch table· Reported by OpenAI· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.47.6%
12GPT 5.2OpenAI GPT-5.4 launch table· Reported by OpenAI· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.40.7%
13GPT 5.2Python · xhigh· Reported by OpenAI· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.40.3%

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.