List your company

Software Engineering Question Answering on SWE-QA Streamlink

8.74ScoreUnpruned
8.4288.5098.598.671May 14, 2026

Evaluation results

MethodLinks
UnprunedBackbone=Claude Opus 4.62026.058.7417.4394.4
LaMRBackbone=Claude Opus 4.62026.058.6818.5353.2
SWE-PrunerBackbone=Claude Opus 4.62026.058.619398.6
SWE-PrunerBackbone=Claude Sonnet 4.52026.058.5923.9587.1
LaMRBackbone=Claude Sonnet 4.52026.058.5316.3467.7
UnprunedBackbone=Claude Sonnet 4.52026.058.4423.4629.3