List your company

Code Reasoning on HumanEval+

97Pass@16ResRL
88.6890.849395.16Oct 16, 2025Nov 17, 2025Dec 20, 2025Jan 22, 2026Feb 24, 2026Mar 29, 2026May 1, 2026

Evaluation results

MethodLinks
ResRLBackbone=Qwen3-4B2026.0597—
NSRBackbone=Qwen3-4B2026.0596.9—
CaSPModel=DeepSeek-R1-Distill-Qwen-7B2025.1096.3283.9
LAD2026.0295.7182.29
GRPOBackbone=Qwen3-4B2026.0595.7—
FlowRLBackbone=Qwen3-4B2026.0595.7—
DAPOBackbone=Qwen3-4B2026.0595.7—
EntAdv2026.0295.0981.05
KLCov2026.0295.0980.75
FlowRL2026.0295.0981.29
ClipCov2026.0294.4981.48
Base ModelModel=DeepSeek-R1-Distill-Qwen-7B2025.1094.4881.02
GRPOModel=DeepSeek-R1-Distill-Qwen-7B2025.1094.4880.25
GRPO2026.0293.2580.9
BackboneBackbone=Qwen3-4B2026.0589—