List your company

Reasoning on BBEH

15.7pass@1DenoiseRL-DAPO
3.6366.7689.913.032Nov 18, 2025Dec 19, 2025Jan 20, 2026Feb 21, 2026Mar 24, 2026Apr 25, 2026May 27, 2026

Evaluation results

MethodLinks
DenoiseRL-DAPOBase Model=Qwen3-4B-Base2026.0515.7
Qwen3-8BBackbone=Qwen3-8B2026.0115.31
DenoiseRL-GRPOBase Model=Qwen3-4B-Base2026.0514.8
RelayLLM (Simple)Student Model Backbone=Qwen3-1.7B2026.0112.67
DenoiseRL-DAPOBase Model=Qwen3-8B-Base2026.0512.6
RelayLLM (Difficulty-Aware)Student Model Backbone=Qwen3-1.7B2026.0112.46
DAPOBase Model=Qwen3-8B-Base2026.0511.7
CITERStudent Model Backbone=Qwen3-1.7B2026.0111.67
DenoiseRL-GRPOBase Model=Qwen3-8B-Base2026.0511.5
GRPOBase Model=Qwen3-4B-Base2026.0511.1
GRPOStudent Model Backbone=Qwen3-1.7B2026.0110.89
GRPOBase Model=Qwen3-8B-Base2026.0510.6
DAPOBase Model=Qwen3-4B-Base2026.0510.4
Base ModelStudent Model Backbone=Qwen3-1.7B2026.019.91
RelayLLM (Difficulty-Aware)Student Model Backbone=Qwen3-0.6B2026.018.56
RelayLLM (Simple)Student Model Backbone=Qwen3-0.6B2026.018.32
CITERStudent Model Backbone=Qwen3-0.6B2026.018.16
GRPOStudent Model Backbone=Qwen3-0.6B2026.017.82
Base ModelStudent Model Backbone=Qwen3-0.6B2026.017.19
GTPOBase Model=Qwen2.5-7B2025.114.4
GRPOBase Model=Qwen2.5-7B2025.114.1
BaseBase Model=Qwen3-4B-Base2026.054.1
BaseBase Model=Qwen3-8B-Base2026.054.1