List your company

Mathematical Reasoning on AMC 23 (pass@16)

95.68pass@16GRPO
90.417691.783893.1594.5162May 7, 2026

Evaluation results

MethodLinks
GRPOBackbone=Qwen3-8B2026.0595.68
GRPO(λ)-0.9Backbone=Qwen3-8B2026.0592.3
S-trace-0.9Backbone=Qwen3-8B2026.0590.62