AA AIME 2025: leaderboard

Artificial Analysis independent evaluation of all 30 problems from the 2025 American Invitational Mathematics Examination (olympiad-level).

Metric: Accuracy (%). Source: artificialanalysis.ai. Saturation forecast: Estimated already saturated. 270 models tracked.

Top models

#ModelScoreOverall rank
1Best Expert100
2GPT-5.2 (xHigh)99#114 (GPT-5.2)
3GPT-5 Codex (High)98.67#130 (GPT-5 Codex)
4GPT-5.2 (Medium)96.67#114 (GPT-5.2)
5DeepSeek V3.2 Speciale96.67#138
6MiMo-V2-Flash (Reasoning)96.33#378 (MiMo-V2-Flash)
7Gemini 3 Pro (Preview) (High)95.67#72 (Gemini 3 Pro (Preview))
8GPT-5.1 Codex (High)95.67#167 (GPT-5.1 Codex)
9GLM-4.7 (Reasoning)95#202 (GLM-4.7)
10Kimi K2 (Thinking)94.67#255 (Kimi K2)
11KAT-Coder-Pro V194.67
12GPT-5 (High)94.33#106 (GPT-5)
13Nova 2.0 Lite (High)94.33#426 (Nova 2.0 Lite)
14GPT-5.1 (High)94#145 (GPT-5.1)
15GPT-OSS-120B (High)93.44#346 (GPT-OSS-120B)

No result here: #3 Claude Opus 5.5, #5 GPT-6 Astra, #6 GPT-6.1 Sol.

Interactive version: theaggregate.ai/benchmark?slug=aa-aime-2025 · How It Works · Data refreshed daily, snapshot 2026-11-06.