Skip to content
Weights Not Walls
Menu

FrontierMath-Tiers-1-3-v2-Private

Re-run by Epoch AI

Tiers 1 to 3 of FrontierMath (second version): unpublished mathematics problems written for Epoch AI, from advanced undergraduate to research level.

Kind
Run independently by Epoch AI, with public logs. Our own re-runs are not built yet.
Score shown
Epoch's column "Best score (across scorers)", as a percentage, higher is better; random guessing scores about 0.0%. Error bars as the source gives them.
Results
127 in all: 26 for open-weight models, 97 for closed ones, 4 where Epoch does not say. 25 linked to a model page here.

Best open-weight result: moonshotai/Kimi-K3.1-Pro, 78.2%, number 25 of 127 results.

FrontierMath-Tiers-1-3-v2-Private results
#ModelWeightsSettingScoreDateSource
1claude-fable-5-2_maxAnthropicclosedeffort max95.1%±1.36 Nov 2026Epoch AI
2gpt-6-astra_maxOpenAIclosedeffort max93.7%±1.430 Aug 2026Epoch AI
3gpt-6.1-sol_maxOpenAIclosedeffort max93.7%±1.429 Sep 2026Epoch AI
4gpt-6.2-astra_maxOpenAIclosedeffort max92.6%±1.627 Oct 2026Epoch AI
5claude-opus-5-5_maxAnthropicclosedeffort max91.2%±1.722 Sep 2026Epoch AI
6gpt-6.1-terra_maxOpenAIclosedeffort max90.9%±1.726 Oct 2026Epoch AI
7gemini-4-carbon_highGoogle DeepMindclosedeffort high90.5%±1.72 Nov 2026Epoch AI
8claude-fable-5-1_maxAnthropicclosedeffort max90.2%±1.81 Sep 2026Epoch AI
9gpt-6-sol_maxOpenAIclosedeffort max89.8%±1.822 Sep 2026Epoch AI
10gpt-5.6-sol_maxOpenAIclosedeffort max89.1%±1.89 Jul 2026Epoch AI log
11claude-sonnet-5-5_maxAnthropicclosedeffort max88.8%±1.929 Sep 2026Epoch AI
12gemini-4-hydrogen_xhighGoogle DeepMindclosedeffort xhigh88.1%±1.913 Nov 2026Epoch AI
13gpt-5.5-pro_xhighOpenAIclosedeffort xhigh87.7%±1.912 Jun 2026Epoch AI
14claude-fable-5_maxAnthropicclosedeffort max87.0%±2.09 Jun 2026Epoch AI log
15gemini-4-argon_highGoogle DeepMindclosedeffort high86.0%±2.113 Oct 2026Epoch AI
16gpt-5.6-terra_maxOpenAIclosedeffort max86.0%±2.19 Jul 2026Epoch AI log
17claude-opus-5_maxAnthropicclosedeffort max85.6%±2.124 Jul 2026Epoch AI log
18gpt-5.5_xhighOpenAIclosedeffort xhigh85.3%±2.111 Jun 2026Epoch AI
19gpt-5.4-pro-2026-03-05_xhighOpenAIclosedeffort xhigh82.5%±2.313 Jun 2026Epoch AI log
20gpt-5.6-luna_maxOpenAIclosedeffort max82.1%±2.39 Jul 2026Epoch AI log
21muse-wonder-1.0_maxMeta AIclosedeffort max82.1%±2.312 Oct 2026Epoch AI
22claude-opus-4-8_maxAnthropicclosedeffort max80.0%±2.410 Jun 2026Epoch AI log
23gpt-6-luna_maxOpenAIclosedeffort max78.9%±2.422 Sep 2026Epoch AI
24gpt-5.4-2026-03-05_xhighOpenAIclosedeffort xhigh78.6%±2.411 Jun 2026Epoch AI log
25moonshotai/Kimi-K3.1-ProEpoch: kimi-k3.1-pro-together_max, Moonshotopeneffort max78.2%±2.46 Nov 2026Epoch AI
26claude-haiku-5-5_maxAnthropicclosedeffort max75.1%±2.67 Oct 2026Epoch AI
27qwen3.8-max_xhighAlibabaclosedeffort xhigh74.7%±2.64 Aug 2026Epoch AI log
28muse-spark-1.3_xhighMeta AIclosedeffort xhigh74.4%±2.616 Sep 2026Epoch AI
29muse-spark-1.3_maxMeta AIclosedeffort max74.0%±2.618 Sep 2026Epoch AI
30gpt-5.2-pro-2025-12-11_xhighOpenAIclosedeffort xhigh74.0%±2.613 Jun 2026Epoch AI
31moonshotai/Kimi-K3Epoch: kimi-k3_max, Moonshotopeneffort max72.2%±2.717 Jul 2026Epoch AI
32gemini-3.7-flash_highGoogle DeepMindclosedeffort high71.6%±2.714 Aug 2026Epoch AI log
33deepseek-ai/DeepSeek-V4.1-Flash-1017Epoch: deepseek-v4.1-flash-1017_max, DeepSeekopeneffort max71.2%±2.719 Oct 2026Epoch AI
34claude-opus-4-7_maxAnthropicclosedeffort max70.2%±2.710 Jun 2026Epoch AI log
35meta-models/Muse-Spark-Open-816B-A42BEpoch: Muse-Spark-Open-816B-A42B_xhigh, Meta AIopeneffort xhigh69.8%±2.715 Oct 2026Epoch AI
36zai-org/GLM-5.3Epoch: glm-5.3_max, Z.ai (Zhipu AI)openeffort max68.8%±2.725 Aug 2026Epoch AI
37gemini-3.8-flash_highGoogle DeepMindclosedeffort high68.4%±2.82 Sep 2026Epoch AI
38gpt-5.2-2025-12-11_xhighOpenAIclosedeffort xhigh67.4%±2.811 Jun 2026Epoch AI
39deepseek-ai/DeepSeek-V4.1-FlashEpoch: deepseek-v4.1-flash_max, DeepSeekopeneffort max67.4%±2.88 Oct 2026Epoch AI
40claude-opus-4-6_maxAnthropicclosedeffort max66.0%±2.811 Jun 2026Epoch AI log
41grok-4.6_xhighxAIclosedeffort xhigh66.0%±2.814 Aug 2026Epoch AI log
42claude-sonnet-5_maxAnthropicclosedeffort max65.6%±2.830 Jun 2026Epoch AI log
43qwen3.8-max-0902_xhighAlibabaclosedeffort xhigh65.6%±2.82 Sep 2026Epoch AI
44deepseek-ai/DeepSeek-V4-Pro-0813Epoch: deepseek-v4-pro-0813_max, DeepSeekopeneffort max64.6%±2.819 Aug 2026Epoch AI log
45qwen3.7-maxAlibabacloseddefault64.6%±2.813 Jun 2026Epoch AI log
46gemini-3.5-flash_highGoogle DeepMindclosedeffort high62.8%±2.910 Jun 2026Epoch AI log
47gemini-3.1-pro-previewGoogle DeepMindcloseddefault59.6%±2.911 Jun 2026Epoch AI log
48zai-org/GLM-5.2Epoch: glm-5.2_max, Z.ai (Zhipu AI)openeffort max59.2%±3.019 Jun 2026Epoch AI
49gemini-3.6-flash_highGoogle DeepMindclosedeffort high58.9%±2.92 Aug 2026Epoch AI log
50mistral-large-4-0_highMistral AIno page matchopeneffort high58.6%±2.912 Oct 2026Epoch AI
51deepseek-ai/DeepSeek-V4-Flash-0731Epoch: deepseek-v4-flash-0731_max, DeepSeekopeneffort max57.5%±2.92 Aug 2026Epoch AI log
52grok-4.5_highxAIclosedeffort high57.2%±2.99 Jul 2026Epoch AI
53moonshotai/Kimi-K2.6Epoch: kimi-k2.6, Moonshotopendefault57.2%±2.910 Jun 2026Epoch AI log
54zai-org/GLM-5.3-FlashEpoch: glm-5.3-flash_max, Z.ai (Zhipu AI)openeffort max55.8%±2.927 Aug 2026Epoch AI
55gpt-5-pro-2025-10-06_highOpenAIclosedeffort high55.8%±2.912 Jun 2026Epoch AI log
56gpt-5-2025-08-07_highOpenAIclosedeffort high55.4%±2.910 Jun 2026Epoch AI
57zai-org/GLM-5.2Epoch: glm-5.2_low, Z.ai (Zhipu AI)openeffort low54.7%±3.029 Aug 2026Epoch AI log
58moonshotai/Kimi-K2.7-CodeEpoch: kimi-k2.7-code, Moonshotopendefault54.0%±3.013 Jun 2026Epoch AI log
59grok-4.7_xhighxAIclosedeffort xhigh53.0%±3.022 Sep 2026Epoch AI
60gemini-3-flash-previewGoogle DeepMindcloseddefault51.2%±3.011 Jun 2026Epoch AI log
61gpt-5.4-mini-2026-03-17_xhighOpenAIclosedeffort xhigh51.2%±3.012 Jun 2026Epoch AI
62gpt-5-mini-2025-08-07_highOpenAIclosedeffort high46.7%±3.012 Jun 2026Epoch AI log
63thinkingmachines/Inkling-SmallEpoch: inkling-small_xhigh, Thinking Machinesopeneffort xhigh46.3%±3.014 Aug 2026Epoch AI log
64deepseek-ai/DeepSeek-V4-ProEpoch: deepseek-v4-pro_max, DeepSeekopeneffort max45.3%±3.017 Jun 2026Epoch AI log
65gpt-5.4-nano-2026-03-17_highOpenAIclosedeffort high44.9%±3.012 Jun 2026Epoch AI log
66grok-4.20-0309-reasoningxAIcloseddefault44.9%±3.013 Jul 2026Epoch AI log
67grok-4.3_highxAIclosedeffort high42.8%±2.917 Jun 2026Epoch AI log
68zai-org/GLM-5.2Epoch: glm-5.2_none, Z.ai (Zhipu AI)openeffort none42.5%±2.929 Aug 2026Epoch AI log
69gpt-5.6-luna_lowOpenAIclosedeffort low41.4%±2.929 Aug 2026Epoch AI log
70gpt-5.6-luna_noneOpenAIclosedeffort none39.6%±2.929 Aug 2026Epoch AI log
71qwen3.6-plusAlibabacloseddefault38.2%±2.930 Aug 2026Epoch AI log
72gpt-5-2025-08-07_lowOpenAIclosedeffort low37.2%±2.927 Aug 2026Epoch AI log
73zai-org/GLM-5.1Epoch: glm-5.1, Z.ai (Zhipu AI)opendefault36.8%±2.929 Aug 2026Epoch AI log
74o4-mini-2025-04-16_highOpenAIclosedeffort high36.1%±2.911 Jun 2026Epoch AI log
75Qwen/Qwen3.6-27BEpoch: qwen3.6-27b, Alibabaopendefault35.1%±2.828 Aug 2026Epoch AI log
76claude-opus-4-5-20251101_32KAnthropiccloseddefault34.4%±2.811 Jun 2026Epoch AI log
77qwen3.7-plus_noneAlibabaclosedeffort none34.4%±2.829 Aug 2026Epoch AI log
78Qwen/Qwen3.6-27BEpoch: qwen3.6-27b_none, Alibabaopeneffort none34.0%±2.829 Aug 2026Epoch AI log
79thinkingmachines/InklingEpoch: Inkling_xhigh, Thinking Machinesopeneffort xhigh33.3%±2.86 Aug 2026Epoch AI log
80o3-2025-04-16_highOpenAIclosedeffort high33.3%±2.827 Aug 2026Epoch AI log
81qwen3.6-plus_noneAlibabaclosedeffort none32.3%±2.829 Aug 2026Epoch AI log
82Qwen/Qwen3.5-397B-A17BEpoch: qwen3.5-397b-a17b_none, Alibabaopeneffort none31.2%±2.730 Aug 2026Epoch AI log
83o3-2025-04-16_mediumOpenAIclosedeffort medium29.8%±2.727 Aug 2026Epoch AI log
84Qwen/Qwen3.5-397B-A17BEpoch: qwen3.5-397b-a17b, Alibabaopendefault29.5%±2.728 Aug 2026Epoch AI log
85o4-mini-2025-04-16_mediumOpenAIclosedeffort medium28.8%±2.727 Aug 2026Epoch AI log
86gemini-3.1-flash-lite_highGoogleclosedeffort high27.7%±2.730 Aug 2026Epoch AI log
87gpt-5.5-instantOpenAIcloseddefault26.3%±2.62 Aug 2026Epoch AI log
88gemini-3.5-flash-lite_highGoogle DeepMindclosedeffort high26.0%±2.62 Aug 2026Epoch AI log
89zai-org/GLM-5.1Epoch: glm-5.1_none, Z.ai (Zhipu AI)openeffort none24.9%±2.628 Aug 2026Epoch AI log
90gemini-2.5-proGoogle DeepMindcloseddefault24.6%±2.611 Jun 2026Epoch AI log
91gpt-5.4-mini-2026-03-17_lowOpenAIclosedeffort low24.6%±2.628 Aug 2026Epoch AI log
92claude-sonnet-4-5-20250929_32KAnthropiccloseddefault23.9%±2.511 Jun 2026Epoch AI log
93gemini-3.1-flash-lite_lowGoogleclosedeffort low22.5%±2.530 Aug 2026Epoch AI log
94qwen3.6-flash_noneAlibabanot statedeffort none22.5%±2.529 Aug 2026Epoch AI log
95gemini-3.1-flash-lite_minimalGoogleclosedeffort minimal21.4%±2.428 Aug 2026Epoch AI log
96gpt-5.4-nano-2026-03-17_lowOpenAIclosedeffort low20.4%±2.428 Aug 2026Epoch AI log
97Qwen/Qwen3.6-35B-A3BEpoch: qwen3.6-35b-a3b_none, Alibabaopeneffort none20.4%±2.430 Aug 2026Epoch AI log
98gpt-5-nano-2025-08-07_highOpenAIclosedeffort high20.0%±2.412 Jun 2026Epoch AI log
99o3-2025-04-16_lowOpenAIclosedeffort low19.3%±2.327 Aug 2026Epoch AI log
100qwen3.7-flashAlibabacloseddefault19.3%±2.329 Aug 2026Epoch AI log
101qwen3.7-flash_noneAlibabaclosedeffort none19.3%±2.329 Aug 2026Epoch AI log
102qwen3-max-2025-09-23Alibabacloseddefault18.9%±2.330 Aug 2026Epoch AI log
103o3-mini-2025-01-31_highOpenAIclosedeffort high18.6%±2.311 Jun 2026Epoch AI log
104gpt-5-2025-08-07_minimalOpenAIclosedeffort minimal18.2%±2.327 Aug 2026Epoch AI log
105gpt-5-mini-2025-08-07_lowOpenAIclosedeffort low18.2%±2.327 Aug 2026Epoch AI log
106qwen3.5-flash_noneAlibabanot statedeffort none18.2%±2.328 Aug 2026Epoch AI log
107Qwen/Qwen3.6-35B-A3BEpoch: qwen3.6-35b-a3b, Alibabaopendefault17.5%±2.329 Aug 2026Epoch AI log
108gpt-5.4-mini-2026-03-17_noneOpenAIclosedeffort none17.2%±2.228 Aug 2026Epoch AI log
109qwen3.6-flashAlibabanot stateddefault17.2%±2.229 Aug 2026Epoch AI log
110o4-mini-2025-04-16_lowOpenAIclosedeffort low16.1%±2.227 Aug 2026Epoch AI log
111o1-2024-12-17_highOpenAIclosedeffort high14.7%±2.128 Aug 2026Epoch AI log
112claude-opus-4-1-20250805_32KAnthropiccloseddefault12.6%±2.011 Jun 2026Epoch AI log
113o3-mini-2025-01-31_mediumOpenAIclosedeffort medium10.5%±1.827 Aug 2026Epoch AI log
114o1-2024-12-17_mediumOpenAIclosedeffort medium10.2%±1.827 Aug 2026Epoch AI log
115qwen3.5-flashAlibabanot stateddefault9.5%±1.728 Aug 2026Epoch AI log
116o1-2024-12-17_lowOpenAIclosedeffort low8.4%±1.627 Aug 2026Epoch AI log
117gpt-4.1-mini-2025-04-14OpenAIcloseddefault6.7%±1.527 Aug 2026Epoch AI log
118gpt-4.1-2025-04-14OpenAIcloseddefault6.0%±1.427 Aug 2026Epoch AI log
119gpt-5-mini-2025-08-07_minimalOpenAIclosedeffort minimal6.0%±1.427 Aug 2026Epoch AI log
120gpt-5-nano-2025-08-07_lowOpenAIclosedeffort low6.0%±1.427 Aug 2026Epoch AI log
121gpt-5.4-nano-2026-03-17_noneOpenAIclosedeffort none4.6%±1.228 Aug 2026Epoch AI log
122o3-mini-2025-01-31_lowOpenAIclosedeffort low3.9%±1.127 Aug 2026Epoch AI log
123gpt-5-nano-2025-08-07_minimalOpenAIclosedeffort minimal1.8%±0.827 Aug 2026Epoch AI log
124gpt-4-turbo-2024-04-09OpenAIcloseddefault0.7%±0.527 Aug 2026Epoch AI log
125gpt-4o-mini-2024-07-18OpenAIcloseddefault0.7%±0.527 Aug 2026Epoch AI log
126gpt-4o-2024-08-06OpenAIcloseddefault0.4%±0.427 Aug 2026Epoch AI log
127gpt-3.5-turbo-0125OpenAIcloseddefault0.0%27 Aug 2026Epoch AI log

Open-weight rows are shaded. Dates are when the result was run or evaluated, where the data says. Data from Epoch AI (CC-BY 4.0), fetched 19 Nov 2026: Epoch AI, ‘Capabilities & benchmarking’. Published online at epoch.ai. Retrieved from ‘https://epoch.ai/benchmarks’ [online resource]. Epoch AI. Method