FrontierMath Tier 4 (2025 set)

Earlier Tier 4 set, superseded by v2 and counted as the same benchmark.

CategoryMath
MetricAccuracy
SourceEpoch AI (evaluator)
LicenseCC BY 4.0
Benchmark pageepoch.ai
StatusScored
Weight1.2 within Math
FitDiscrimination 2.20, noise 0.59
LineageShares weight with frontiermath_tier_4

Results

#ModelProviderResultDate
1 AI Co-Mathematician Google 47.9% 2026-05-08
2 GPT-5.5 Pro Extra High OpenAI 39.6% 2026-04-23
3 GPT-5.5 Pro High OpenAI 39.6% 2026-04-23
4 GPT-5.4 Pro OpenAI 37.5% 2026-03-06
5 GPT-5.5 Extra High OpenAI 35.4% 2026-04-23
6 GPT-5.2 Pro OpenAI 31.3% 2025-12-24
7 Claude Opus 4.8 Max Anthropic 31.3% 2026-06-08
8 GPT-5.4 Extra High OpenAI 27.1% 2026-03-06
9 Claude Opus 4.7 Extra High Anthropic 22.9% 2026-04-17
10 Claude Opus 4.6 Max Anthropic 22.9% 2026-02-12
11 Claude Opus 4.6 Thinking Anthropic 20.8% 2026-02-06
12 GPT-5.2 Extra High OpenAI 18.8% 2025-12-14
13 GPT-5.2 High OpenAI 18.8% 2025-12-11
14 Gemini 3 Pro Preview Google 18.8% 2025-11-21
15 Gemini 3.1 Pro Preview Google 16.7% 2026-02-19
16 GPT-5.2 Medium OpenAI 16.7% 2025-12-11
17 Muse Spark Meta 14.6% 2026-04-08
18 GPT-5 Pro High OpenAI 14.6% 2025-10-09
19 Kimi K2.6 Moonshot AI 14.6% 2026-05-08
20 Claude Opus 4.6 Anthropic 14.6% 2026-02-05
21 Gemini 3.5 Flash High Google 14.6% 2026-05-25
22 GLM-5.1 Zhipu AI 12.5% 2026-05-12
23 GPT-5.1 High OpenAI 12.5% 2025-11-13
24 GPT-5 High OpenAI 12.5% 2025-10-30
25 Gemini 2.5 Deep Think Google 10.4% 2026-11-11
26 Qwen3.6 Plus Alibaba 8.3% 2026-05-12
27 Claude Sonnet 4.6 Thinking Anthropic 8.3% 2026-02-21
28 GPT-5.4 nano High OpenAI 6.3% 2026-04-15
29 GPT-5.2 Low OpenAI 6.3% 2025-12-11
30 GPT-5 Mini High OpenAI 6.3% 2025-10-30
31 GPT-5 Medium OpenAI 6.3% 2025-08-07
32 o4-mini High OpenAI 6.3% 2025-07-01
33 Kimi K2.5 Moonshot AI 4.2% 2026-02-02
34 Claude Opus 4.5 Anthropic 4.2% 2025-11-25
35 Gemini 2.5 Pro Google 4.2% 2025-07-03
36 Gemini 2.5 Flash Google 4.2% 2025-12-18
37 Gemini 3 Flash Preview Google 4.2% 2025-12-17
38 Qwen3.6 Max Preview Alibaba 4.2% 2026-05-28
39 Claude Opus 4.5 Thinking Anthropic 4.2% 2025-11-25
40 GPT-5.1 Medium OpenAI 4.2% 2025-11-17
41 Claude Sonnet 4.5 Thinking Anthropic 4.2% 2025-10-22
42 GPT-5 Mini Medium OpenAI 4.2% 2025-08-07
43 Claude Opus 4.1 Thinking Anthropic 4.2% 2025-08-05
44 Claude Opus 4 Thinking Anthropic 4.2% 2025-07-01
45 o3-mini High OpenAI 4.2% 2025-07-01
46 GLM-4.6 Zhipu AI 2.1% 2025-12-08
47 GLM-5 Zhipu AI 2.1% 2026-02-19
48 DeepSeek V3.2 DeepSeek 2.1% 2025-12-16
49 Claude Sonnet 4.5 Anthropic 2.1% 2025-09-29
50 Gemini 2.5 Pro Preview 06-05 Google 2.1% 2025-07-03
51 Qwen3.5 Plus Alibaba 2.1% 2026-05-15
52 Grok 4 xAI 2.1% 2025-08-11
53 Grok 4 Heavy xAI 2.1% 2025-10-09
54 o4-mini Medium OpenAI 2.1% 2025-08-07
55 GPT-5.4 mini High OpenAI 2.1% 2026-04-15
56 Claude Haiku 4.5 Thinking Anthropic 2.1% 2025-10-22
57 GPT-5 Nano Medium OpenAI 2.1% 2025-08-07
58 o3 High OpenAI 2.1% 2025-07-01
59 Kimi K2 Thinking Moonshot AI 0.0% 2025-12-04
60 Claude Opus 4 Anthropic 0.0% 2025-07-01
61 Claude Sonnet 4 Anthropic 0.0% 2025-07-01
62 Claude Sonnet 3.5 v2 Anthropic 0.0% 2025-07-01
63 GLM-4.7 Zhipu AI 0.0% 2026-01-30
64 GPT-4.1 OpenAI 0.0% 2025-07-01
65 Qwen3 235B-A22B Thinking 2507 Alibaba 0.0% 2025-12-11
66 Qwen3.6 Flash Alibaba 0.0% 2026-05-12
67 Qwen3.5 Flash Alibaba 0.0% 2026-05-12
68 Claude 3.5 Sonnet Anthropic 0.0% 2025-07-01
69 Grok 3 xAI 0.0% 2025-07-01
70 GPT-5 Nano High OpenAI 0.0% 2025-10-30