Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

Mathematical Reasoning on AIME 25 (Throughput/Speedup)

9,043Throughput (tokens/s)

Draft-OPD

76.122,404.064,7327,059.94May 28, 2026
Updated 5d ago

Evaluation Results

MethodLinks
2026.05
9,04386.59
2026.05
8,410-6.06
2026.05
7,29776.59
2026.05
6,841-6.06
2026.05
6,645116.42
2026.05
5,985-5.67
2026.05
5,729166.42
2026.05
5,07176.59
2026.05
4,956-5.67
2026.05
4,750-6.06
2026.05
4,718175.32
2026.05
4,127146.42
2026.05
4,014-4.54
2026.05
3,612-5.67
2026.05
3,187165.32
2026.05
2,98486.59
2026.05
2,755-6.06
2026.05
2,738-4.54
2026.05
2,465166.42
2026.05
2,121-5.67
2026.05
2,111135.32
2026.05
1,858-4.54
2026.05
1,229135.32
2026.05
1,086-4.54
2026.05
96966.59
2026.05
912-6.06
2026.05
741126.42
2026.05
662-5.67
2026.05
476135.32
2026.05
421-4.54