FrontierMath legacy aggregate

Legacy FrontierMath values retained for historical model pages. This field is not used in current rankings because it can mix prior benchmark versions and slices.

7 models

Higher is better.

1GPT-5.6 SolOpenAI89.0
2GPT-5.6 TerraOpenAI84.9
3GPT-5.6 LunaOpenAI78.6
4GPT-5.5 ProOpenAI52.4
5GPT-5.5OpenAI51.7
6GPT-5.4 ProOpenAI50.0
7Claude Opus 4.7 (Adaptive)Anthropic43.8

Higher is better. Bar colour: blue proprietary, orange open weights, green other.

About this benchmark

Metric: Open-ended mathematical reasoning with tool access
Tasks: Historical aggregate
Authors: Epoch AI (2024)
Source: FrontierMath: A Benchmark for Evaluating Advanced Mathematical Reasoning in AI ↗

Scores are as reported by model developers or evaluators and may use different settings (for example reasoning effort or tools). Compare with care.