Benchmark

SWE-MM

SWE-MM evaluates software-engineering agents on repository tasks that require understanding both source code and visual evidence.

Modality
multimodal
Categories
multimodal, agents, code, vision
Openness
unknown
Source
llm_stats
Reported scores
1

Reported scores

llm_stats

ModelOrganizationReported valueReportedEvidence
Qwen3.8-27BQwen0.3862026-08-14evidence

Scores are partitioned by the source that reported them and are never merged into a single cross-source ranking, because the sources measure different things and say so.