AIME 2025

Artificial Analysis · 50 models

Rank	Model	Score
1	GPT-5.2 (Xhigh)	0.99
2	GPT-5 Codex (High)	0.9867
3	Gemini 3 Flash Preview (Reasoning)	0.97
4	DeepSeek V3.2 Speciale	0.9667
4	GPT-5.2 (Medium)	0.9667
6	MiMo-V2-Flash (Reasoning)	0.9633
7	GPT-5.1 Codex (High)	0.9567
7	Gemini 3 Pro Preview (High)	0.9567
9	GLM-4.7 (Reasoning)	0.95
10	KAT-Coder-Pro V1	0.9467
10	Kimi K2 Thinking	0.9467
12	GPT-5 (High)	0.9433
12	Nova 2.0 Lite (High)	0.9433
14	GPT-5.1 (High)	0.94
15	gpt-oss-120b (High)	0.9344
16	Grok 4	0.9267
17	DeepSeek V3.2 (Reasoning)	0.92
18	GPT-5 (Medium)	0.9167
18	GPT-5.1 Codex mini (High)	0.9167
20	Claude Opus 4.5 (Reasoning)	0.9133
21	NVIDIA Nemotron 3 Nano 30B A3B (Reasoning)	0.91
21	Qwen3 235B A22B 2507 (Reasoning)	0.91
23	GPT-5 mini (High)	0.9067
23	o4-mini (High)	0.9067
25	K-EXAONE (Reasoning)	0.9033
26	DeepSeek V3.1 (Reasoning)	0.8967
26	DeepSeek V3.1 Terminus (Reasoning)	0.8967
26	Grok 4 Fast (Reasoning)	0.8967
26	Nova 2.0 Omni (Medium)	0.8967
30	Grok 4.1 Fast (Reasoning)	0.8933
30	Ring-1T	0.8933
30	gpt-oss-20b (High)	0.8933
33	Nova 2.0 Pro Preview (Medium)	0.89
34	Nova 2.0 Lite (Medium)	0.8867
35	Qwen3 VL 235B A22B (Reasoning)	0.8833
35	o3	0.8833
37	Apriel-v1.6-15B-Thinker	0.88
37	Claude 4.5 Sonnet (Reasoning)	0.88
37	INTELLECT-3 (Based on GLM-4.5-Air)	0.88
40	DeepSeek V3.2 Exp (Reasoning)	0.8767
40	Gemini 2.5 Pro	0.8767
42	Apriel-v1.5-15B-Thinker	0.875
43	Gemini 3 Pro Preview (Low)	0.8667
44	GLM-4.6 (Reasoning)	0.86
45	GLM-4.6V (Reasoning)	0.8533
46	ERNIE 5.0 Thinking Preview	0.85
46	GPT-5 mini (Medium)	0.85
48	Grok 3 mini Reasoning (High)	0.8467
48	Qwen3 VL 32B (Reasoning)	0.8467
48	Seed-OSS-36B-Instruct	0.8467