Benchmark

DeepSWE

DeepSWE is a software engineering agent benchmark evaluated with the mini-swe-agent harness, where each task is solved in an isolated container with no…

Modality
text
Categories
agents, code
Openness
unknown
Source
llm_stats
Reported scores
11

Reported scores

llm_stats

ModelOrganizationReported valueReportedEvidence
DeepSeek-V4-Flash-0731DeepSeek0.5442026-07-31evidence
DeepSeek-V4-Pro-0813DeepSeek0.6272026-08-13evidence
GLM-5.2Z.ai0.4622026-06-16evidence
GPT-5.6 LunaOpenAI0.6722026-07-09evidence
GPT-5.6 SolOpenAI0.7272026-07-09evidence
GPT-5.6 TerraOpenAI0.6962026-07-09evidence
Grok 4.5xAI0.532026-07-16evidence
Hy3Tencent0.282026-07-06evidence
Kimi K3Moonshot AI0.6752026-07-16evidence
Seed 2.1 ProByteDance0.3272026-06-24evidence
Seed 2.1 TurboByteDance0.232026-06-24evidence

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.