List your company

Code Generation on LiveCodeBench Medium

96.79AccuracyWait
9.159631.909854.6677.4102Sep 26, 2025

Evaluation results

MethodLinks
WaitModel=Gemini 3 Pro, Strategy=Wait2025.0996.79—
HEARTModel=Gemini 3 Pro, Strategy=HEART2025.0996.78—
HEARTModel=Gemini 3 Flash, Strategy=HEART2025.0996.53—
Self ReflectionModel=Gemini 3 Pro, Strategy=Self Reflection2025.0996.53—
CoTModel=Gemini 3 Pro, Strategy=CoT2025.0996.52—
Self ReflectionModel=Gemini 3 Flash, Strategy=Self Reflection2025.0996—
CoTModel=Gemini 3 Flash, Strategy=CoT2025.0995.73—
WaitModel=Gemini 3 Flash, Strategy=Wait2025.0995.73—
HEARTModel=Gemini 2.5 Pro2025.0984.57—
HEARTModel=Gemini 2.5 Flash2025.0984.37—
WaitModel=Gemini 2.5 Pro2025.0983.02—
CoTModel=Gemini 2.5 Pro2025.0982.47—
VanillaModel=Gemini 2.5 Flash2025.0982.25—
Self ReflectionModel=Gemini 2.5 Pro2025.0982.14—
Self ReflectionModel=Gemini 2.5 Flash2025.0981.35—
VanillaModel=Gemini 2.5 Pro2025.0981.2—
WaitModel=Gemini 2.5 Flash2025.0980.17—
CoTModel=Gemini 2.5 Flash2025.0980.02—
HEARTModel=Gemma3-12b-Instruct2025.0913.32—
Self ReflectionModel=Gemma3-12b-Instruct2025.0913.05—
CoTModel=Gemma3-12b-Instruct2025.0913.05—
WaitModel=Gemma3-12b-Instruct2025.0913.05—
VanillaModel=Gemma3-12b-Instruct2025.0912.53—
DeepSeek-R1Evaluation Protocol=Leaderboard2025.12—77.7
DEPO-iter3Base Model=Qwen2.5-Coder-7B-Instruct, Iteration=32026.05—18.8
DPO-iter3Base Model=Qwen2.5-Coder-7B-Instruct, Iteration=32026.05—14.5
DreamPRM-CodeEvaluation Protocol=Test-Time Scaling, Base Model=O4-mini-high2025.12—92.3
Gemini-2.5Evaluation Protocol=Leaderboard2025.12—82.1
O3Evaluation Protocol=Leaderboard2025.12—71.8
O4-mini-highEvaluation Protocol=Leaderboard2025.12—89.7
ORMEvaluation Protocol=Test-Time Scaling, Base Model=O4-mini-high2025.12—89.7
POPO-iter3Base Model=Qwen2.5-Coder-7B-Instruct, Iteration=32026.05—15
Qwen2.5-Coder-7B-Instruct2026.05—3.8
XPO-iter3Base Model=Qwen2.5-Coder-7B-Instruct, Iteration=32026.05—15.2