List your company

Software Engineering on PaperBench Code (dev)

78.2ScoreSTORM-Combined
46.4854.71562.9571.185May 19, 2026

Evaluation results

MethodLinks
STORM-CombinedModel=Claude Sonnet 4.6, Protocol=STORM-Combined2026.0578.217.331.4
GitWorktree-CombinedModel=Claude Sonnet 4.6, Protocol=GitWorktree-Combined2026.0576.622.730.8
STORMModel=Claude Sonnet 4.6, Protocol=STORM2026.0574.112.624.1
GitWorktreeModel=Claude Sonnet 4.6, Protocol=GitWorktree2026.0572.717.222.5
Single-AgentModel=Claude Sonnet 4.6, Protocol=Single-Agent2026.0568.712.511.6
STORM-CombinedModel=DeepSeek V4 Pro, Protocol=STORM-Combined2026.0568.310.941.7
STORMModel=DeepSeek V4 Pro, Protocol=STORM2026.0566.59.735.5
Single-AgentModel=DeepSeek V4 Pro, Protocol=Single-Agent2026.0562.94.116.5
GitWorktree-CombinedModel=DeepSeek V4 Pro, Protocol=GitWorktree-Combined2026.0560.89.141.3
STORM-CombinedModel=Qwen 3.6 Plus, Protocol=STORM-Combined2026.055710.729.6
GitWorktreeModel=DeepSeek V4 Pro, Protocol=GitWorktree2026.0555.88.835.7
GitWorktree-CombinedModel=Qwen 3.6 Plus, Protocol=GitWorktree-Combined2026.0555.412.143.1
STORMModel=Qwen 3.6 Plus, Protocol=STORM2026.05558.221
GitWorktreeModel=Qwen 3.6 Plus, Protocol=GitWorktree2026.0551.613.436
Single-AgentModel=Qwen 3.6 Plus, Protocol=Single-Agent2026.0547.74.916.1