Skip to main content
Evaluation sources/FrontierMath v2 Tiers 1–3
ParticipatingReasoning

FrontierMath v2 Tiers 1–3

Epoch AI · Public model evaluation results from FrontierMath v2 Tiers 1–3.

Official evaluation
Source description

This source measures concrete delivery performance in Reasoning; raw scores are shown separately from the normalized capability score.

Limitation: Different versions, reasoning settings, and test sets may change scores; read them with the data date.

How it is used

Participating sources affect capability scores; reference, cross-reference, and observing sources are for comparison and verification only.

Model results from this source

Claude Fable 5.20.9508771972.72026-11-10
GPT-6 Astra0.9368421171.92026-11-05
GPT-6.1 Sol0.9368421171.92026-11-05
Claude Opus 5.50.912280770.42026-11-05
GPT-6.1 Terra0.9087719370.22026-11-05
Gemini 4 Carbon0.9052631669.92026-11-11
Claude Fable 5.10.9017543969.72026-11-05
GPT-6 Sol0.8982456169.52026-11-05
GPT-5.6 Sol0.8912280769.12026-11-05
Claude Sonnet 5.50.887719368.92026-11-05
GPT-5.5 Pro0.8771929868.22026-11-05
Claude Fable 50.8701754467.82026-11-05
GPT-5.6 Terra0.8596491267.12026-11-05
Gemini 4 Argon0.8596491267.12026-10-17
Gemini 4 Argon0.8596491267.12026-11-05
Gemini 4 Argon0.8596491267.12026-10-18
Claude Opus 50.8561403566.92026-11-05
GPT-5.50.8526315866.72026-11-05
GPT-5.4 Pro0.824561465.02026-11-05
GPT-5.6 Luna0.8210526364.82026-11-05
Muse Wonder0.8210526364.82026-11-05
Muse Wonder0.8210526364.82026-10-18
Muse Wonder0.8210526364.82026-10-17
Claude Opus 4.80.863.52026-11-05
GPT-6 Luna0.7894736862.82026-11-05
GPT-5.40.7859649162.62026-11-05
Claude Haiku 5.50.7508771960.52026-11-05
Qwen3.8 Max0.7473684260.32026-11-05
Muse Spark 1.30.7438596560.12026-11-05
GPT-5.2 Pro0.7459.82026-11-05
Kimi K30.7218309958.72026-11-05
DeepSeek V4.1 Flash 10170.712280758.12026-11-05
Claude Opus 4.70.7017543957.52026-11-05
GLM 5.30.687719356.62026-11-05
Gemini 3.8 Flash0.6842105356.42026-11-05
GPT-5.20.67455.82026-11-05
DeepSeek V4.1 Flash0.6736842155.72026-11-05
Claude Opus 4.60.6596491254.92026-11-05
Grok 4.60.6596491254.92026-11-05
Claude Sonnet 50.6561403554.72026-11-05
Qwen3.7 Max0.6456140454.02026-11-05
DeepSeek V4 Pro 08130.6456140454.02026-11-05
Gemini 3.1 Pro Preview0.5964912351.02026-11-05
GLM-5.20.5920577650.72026-11-05
Gemini 3.6 Flash0.5894736850.62026-11-05
Mistral Large 40.5859649150.42026-10-31
Mistral Large 40.5859649150.42026-10-17
Mistral Large 40.5859649150.42026-10-18
DeepSeek V4 Flash 07310.575438649.72026-11-05
Kimi K2.60.5719298249.52026-11-05
Grok 4.50.5719298249.52026-11-05
GPT-5 Pro0.5578947448.72026-11-05
GLM-5.3 Flash0.5578947448.72026-10-18
GLM-5.3 Flash0.5578947448.72026-11-05
GLM-5.3 Flash0.5578947448.72026-10-17
GPT-50.5543859648.42026-11-05
Kimi K2.7 Code0.5403508847.62026-11-05
Grok 4.70.5298245646.92026-11-05
Gemini 3 Flash Preview0.512280745.92026-11-05
GPT-5.4 mini0.512280745.92026-11-05
GPT-5 mini0.4666666743.12026-11-05
Inkling Small0.4631578942.82026-11-05
Grok 4.200.4491228142.02026-11-05
Qwen3.6 Plus0.3824561437.92026-11-05
GLM-5.10.3684210537.02026-11-05
o4 Mini High0.3614035136.62026-11-05
Qwen3.6 27B0.3508771936.02026-11-05
Qwen3.7 Plus0.3438596535.52026-11-05
Claude Opus 4.50.3438596535.52026-11-05
o30.3333333334.92026-11-05
Inkling0.3333333334.92026-11-05
Qwen3.5 397B-A17B0.312280733.62026-11-05
o4-mini0.287719332.12026-11-05
Gemini 3.1 Flash Lite0.2771929831.42026-11-05
GPT-5.5 Instant0.2631578930.62026-11-05
Gemini 3.5 Flash Lite0.2596491230.42026-11-05
Gemini 2.5 Pro0.2456140429.52026-11-05
Claude Sonnet 4.50.2385964929.12026-11-05
Qwen3.6 Flash0.224561428.22026-11-05
Qwen3.6 35B-A3B0.2035087726.92026-11-05
GPT-5 nano0.226.72026-11-05
Qwen3.7 Flash0.1929824626.32026-11-05
Qwen3 Max0.1894736826.12026-11-05
o3 Mini High0.1859649125.92026-11-05
Qwen3.5-Flash0.1824561425.62026-11-05
o10.1473684223.52026-11-03
o3-mini0.1052631620.92026-11-05

Frequently asked questions