List your company

Mathematical Problem Solving on AIME 2025 (Pass@k)

64.7Pass@kGCPO
9.26823.65938.0552.441May 12, 2026

Evaluation results

MethodLinks
GCPOk=64, Base Model=Qwen3-1.7B2026.0564.7
GCPOk=32, Base Model=Qwen3-1.7B2026.0559.8
GCPOk=16, Base Model=Qwen3-1.7B2026.0552.5
DQOk=64, Base Model=Qwen3-1.7B2026.0546.8
GCPOk=8, Base Model=Qwen3-1.7B2026.0544.7
DQOk=32, Base Model=Qwen3-1.7B2026.0542.9
GRPOk=64, Base Model=Qwen3-1.7B2026.0542.4
GRPOk=32, Base Model=Qwen3-1.7B2026.0538.5
GCPOk=4, Base Model=Qwen3-1.7B2026.0538.2
DQOk=16, Base Model=Qwen3-1.7B2026.0537.7
Basek=64, Base Model=Qwen3-1.7B2026.0537.5
DQOk=8, Base Model=Qwen3-1.7B2026.0534.2
GRPOk=16, Base Model=Qwen3-1.7B2026.0534
Basek=32, Base Model=Qwen3-1.7B2026.0531.4
DQOk=4, Base Model=Qwen3-1.7B2026.0531.1
GCPOk=2, Base Model=Qwen3-1.7B2026.0530.8
GRPOk=8, Base Model=Qwen3-1.7B2026.0530.8
GRPOk=4, Base Model=Qwen3-1.7B2026.0528.3
DQOk=2, Base Model=Qwen3-1.7B2026.0526.3
Basek=16, Base Model=Qwen3-1.7B2026.0524.8
GRPOk=2, Base Model=Qwen3-1.7B2026.0523.9
Basek=8, Base Model=Qwen3-1.7B2026.0520.5
Basek=4, Base Model=Qwen3-1.7B2026.0515.7
Basek=2, Base Model=Qwen3-1.7B2026.0511.4