Mathematical Reasoning on AIME 2024 (VeRA-E) benchmark leaderboard · SOTA2 Research

Mathematical Reasoning on AIME 2024 (VeRA-E)

94SeedGemini-3-Pro-Preview

Evaluation Results

Method	Links
Gemini-3-Pro-Preview2026.01		94	81.3	-12.7
GPT-5.1-high2026.01		92	75.3	-16.7
DeepSeek-V3.2-thinking2026.01		92	79	-13
GLM-4.62026.01		90.7	71.7	-19
DeepSeek-V3.1-thinking2026.01		90.7	77.7	-13
GPT-5-high2026.01		90	74.7	-15.3
Seed-1.6-1015-high2026.01		89.3	75	-14.3
Gemini-2.5-Pro2026.01		88	73.3	-14.7
Seed-1.6-Lite-1015-high2026.01		87.3	73	-14.3
Kimi-K2-thinking2026.01		86.7	75.3	-11.3
Seed-1.6-Thinking-07152026.01		85.3	72	-13.3
Minimax-M22026.01		84.7	55.7	-29
Claude-Sonnet-4.5-thinking2026.01		84	74	-10
qwen3-max-09232026.01		82.7	68.7	-14
Kimi-K2-09052026.01		65.3	51.3	-14
GPT-5.1-chat-latest2026.01		48.7	46	-2.7