List your company
ResearchFormerly WizWand

Code Generation on OJBench ICPC 2025 (test)

19.18AccuracyPoT
4.9328.63112.3316.029Jan 28, 2026
Updated 8mo ago

Evaluation results

MethodLinks
PoTBase Model=Qwen3-4B-Instruct-25072026.0119.18
ToTBase Model=Qwen3-4B-Instruct-2507, Method Category=Search-Based Reasoning2026.0116.43
AB-MCTSBase Model=Qwen3-4B-Instruct-2507, Method Category=Search-Based Reasoning2026.0115.07
RethinkMCTSBase Model=Qwen3-4B-Instruct-2507, Method Category=Search-Based Reasoning2026.0115.07
LATSBase Model=Qwen3-4B-Instruct-2507, Method Category=Search-Based Reasoning2026.0113.69
ReflexionBase Model=Qwen3-4B-Instruct-2507, Method Category=Self-Refinement / Debugging2026.0110.59
PG-TDBase Model=Qwen3-4B-Instruct-2507, Method Category=Search-Based Reasoning2026.0110.59
Deepseek-V3Inference Strategy=Zero-shot2026.019.59
Best of N (N=20)Base Model=Qwen3-4B-Instruct-2507, Method Category=Standard Inference2026.019.59
Gemini-2.5-flashInference Strategy=Zero-shot2026.018.22
Few-shot + CoTBase Model=Qwen3-4B-Instruct-2507, Method Category=Standard Inference2026.018.22
CODETBase Model=Qwen3-4B-Instruct-2507, Method Category=Self-Refinement / Debugging2026.018.22
RAPBase Model=Qwen3-4B-Instruct-2507, Method Category=Self-Refinement / Debugging2026.018.22
Gpt-4oInference Strategy=Zero-shot2026.016.8
Claude-Opus-4Inference Strategy=Zero-shot2026.016.8
Zero-shotBase Model=Qwen3-4B-Instruct-2507, Method Category=Standard Inference2026.016.8
Qwen3-235B-A22BInference Strategy=Zero-shot2026.015.48
Self-RefineBase Model=Qwen3-4B-Instruct-2507, Method Category=Self-Refinement / Debugging2026.015.48