benchlm.ai

American Invitational Mathematics Examination 2025 (AIME 2025)

The most recent AIME examination, featuring 15 challenging mathematics problems testing olympiad-level mathematical reasoning with integer answers from 000-999.

About AIME 2025

Year

2025

Tasks

15 problems

Format

Integer answers 000-999

Difficulty

High school olympiad level

AIME 2025 represents the current standard for intermediate-level mathematical olympiad problems. Success requires sophisticated mathematical reasoning and problem-solving techniques.

American Invitational Mathematics Examination

Leaderboard (88 models)

#1GPT-5.4
98
#2Gemini 3.1 Pro
98
#3Claude Opus 4.6
98
#4GPT-5.3 Codex
98
#5Grok 4.1
98
#6GPT-5.2
98
#7GPT-5.2-Codex
98
#8Gemini 3 Pro Deep Think
98
#9Claude Sonnet 4.6
98
#10Claude Opus 4.5
98
#11Gemini 3 Pro
98
#12GPT-5.1-Codex-Max
98
#13GPT-5.1
98
#14GLM-5 (Reasoning)
98
#15Claude Sonnet 4.5
98
#16Grok 4.1 Fast
97
#17GPT-5 (high)
96
#18o1-preview
95
#19Kimi K2.5 (Reasoning)
95
#20GPT-5 (medium)
94
#21Qwen3.5 397B (Reasoning)
94
#22GPT-5 mini
91
#23o3-pro
91
#24o3
89
#25GLM-5
89
#26Grok 4
88
#27DeepSeek V3.2 (Thinking)
88
#28GLM-4.7
87
#29Qwen2.5-1M
86
#30Qwen2.5-72B
85
#31Gemini 2.5 Pro
85
#32DeepSeek V3.2
85
#33o4-mini (high)
84
#34Qwen3.5 397B
84
#35DeepSeek Coder 2.0
82
#36DeepSeek LLM 2.0
81
#37DeepSeekMath V2
81
#38MiMo-V2-Flash
80
#39Kimi K2.5
78
#40Claude 4.1 Opus
77
#41Mistral Large 3
77
#42Nemotron 3 Ultra 500B
75
#43Claude 4 Sonnet
74
#44MiniMax M2.5
74
#45Llama 3.1 405B
71
#46Gemini 3 Flash
71
#47Mistral Large 2
69
#48Claude Haiku 4.5
69
#49GPT-4o
67
#50GLM-4.7-Flash
67
#51Mistral 8x7B
66
#52Claude 3.5 Sonnet
66
#53Nemotron 3 Super 100B
66
#54Gemini 1.5 Pro
65
#55Grok Code Fast 1
65
#56Gemini 3.1 Flash-Lite
64
#57Gemini 1.0 Pro
63
#58Claude 3 Opus
62
#59GPT-4 Turbo
61
#60Llama 3 70B
59
#61Nemotron 3 Nano 30B
58
#62Claude 3 Haiku
57
#63Nemotron-4 15B
55
#64Moonshot v1
54
#65Z-1
53
#66GPT-OSS 120B
52
#67Gemini 2.5 Flash
51
#68Nemotron Ultra 253B
50
#69Llama 4 Behemoth
49
#70Llama 4 Scout
48
#71Llama 4 Maverick
47
#72Gemma 3 27B
46
#73DeepSeek-R1
45
#74Qwen2.5-VL-32B
44
#75Grok 3 [Beta]
43
#76Nova Pro
42
#77Qwen3 235B 2507 (Reasoning)
41
#78Qwen3 235B 2507
40
#79Claude 4.1 Opus Thinking
39
#80GLM-4.5
38
#81MiniMax M1 80k
37
#82GLM-4.5-Air
36
#83DeepSeek V3.1 (Reasoning)
35
#84DeepSeek V3.1
34
#85Kimi K2
33
#86GPT-OSS 20B
32
#87Mistral 7B v0.3
31
#88Mistral 8x7B v0.2
30

FAQ

What does AIME 2025 measure?

The most recent AIME examination, featuring 15 challenging mathematics problems testing olympiad-level mathematical reasoning with integer answers from 000-999.

Which model scores highest on AIME 2025?

GPT-5.4 by OpenAI currently leads with a score of 98 on AIME 2025.

How many models are evaluated on AIME 2025?

88 AI models have been evaluated on AIME 2025 on BenchLM.