List your company
ResearchFormerly WizWand

Code Generation on LiveCodeBench v5

72.1AccuracyBF16
-2.2007217.0888936.378555.66811Jan 27, 2026Feb 19, 2026Mar 14, 2026Apr 7, 2026Apr 30, 2026May 23, 2026Jun 16, 2026
Updated 4mo ago

Evaluation results

MethodLinks
BF16Precision=BF16, Backbone=Nemotron 3 Nano2026.0172.1
NVFP4 PTQPrecision=NVFP4, Type=PTQ, Backbone=Nemotron 3 Nano2026.0168.9
NVFP4 QADPrecision=NVFP4, Type=QAD, Backbone=Nemotron 3 Nano2026.0168.9
NVFP4 QATPrecision=NVFP4, Type=QAT, Backbone=Nemotron 3 Nano2026.0162
DRIFTBase Model=OpenR1-Distill-7B2026.0650.21
IF (Impl. w. GraSS)Base Model=OpenR1-Distill-7B2026.0648.18
QuratingBase Model=OpenR1-Distill-7B2026.0647.63
RDSBase Model=OpenR1-Distill-7B2026.0645.19
Random BaselineBase Model=OpenR1-Distill-7B2026.0644.81
DSIRBase Model=OpenR1-Distill-7B2026.0644.46
BM25Base Model=OpenR1-Distill-7B2026.0644.19
LESSBase Model=OpenR1-Distill-7B2026.0644.1
OpenR1-Distill-7BBase Model=OpenR1-Distill-7B2026.0643.86
Self-DistillationBase Model=OpenR1-Distill-7B2026.0620.31
DRIFTBase Model=Olmo3-7B-Instruct-SFT2026.0616.78
QuratingBase Model=Olmo3-7B-Instruct-SFT2026.0616.52
BM25Base Model=Olmo3-7B-Instruct-SFT2026.0616.28
RDSBase Model=Olmo3-7B-Instruct-SFT2026.0615.55
Olmo3-7B-Instruct-SFTBase Model=Olmo3-7B-Instruct-SFT2026.0615.43
Random BaselineBase Model=Olmo3-7B-Instruct-SFT2026.0615.07
IF (Impl. w. GraSS)Base Model=Olmo3-7B-Instruct-SFT2026.0614.53
LESSBase Model=Olmo3-7B-Instruct-SFT2026.0614.22
DSIRBase Model=Olmo3-7B-Instruct-SFT2026.0613.51
Self-DistillationBase Model=Olmo3-7B-Instruct-SFT2026.069.48
PromptCoT 2.0training_budget=11,000 mixed trajectories, recipe=GRPO, evaluation=Best-of-52026.020.742
Agentic Proposingtraining_budget=11,000 mixed trajectories, recipe=GRPO, evaluation=Best-of-5, proposer=Agentic-Proposer-30B2026.020.734
OpenCodeReasoningtraining_budget=11,000 mixed trajectories, recipe=GRPO, evaluation=Best-of-52026.020.708
OpenThoughts-S3training_budget=11,000 mixed trajectories, recipe=GRPO, evaluation=Best-of-52026.020.683
OpenR1training_budget=11,000 mixed trajectories, recipe=GRPO, evaluation=Best-of-52026.020.683
Qwen3-30B-A3B-Thinking-2507mode=zero-shot, evaluation=Best-of-52026.020.681
OpenMathReasoningtraining_budget=11,000 mixed trajectories, recipe=GRPO, evaluation=Best-of-52026.020.657