Fetching from the wire…
Research2026-08-21 · source-backed
Semantics-preserving transformations of SWE-bench repos produced statistically significant degradation in 6 of 16 model-scaffold-dataset configurations. Rankings flipped by scaffold: Qwen looked robust under mini-SWE agent and brittle under OpenCode. arXiv The simpler scaffold was more robust overall, which argues against assuming a heavier harness buys reliability.
Each link below shares sources, entities, or timing with this story.
OpenCode uses Qwen / Shared entities / Same source domain / Shared topic / Earlier coverage
Linked by a graph relationship (OpenCode uses Qwen); both cover Opencode, SWE; reported by the same outlet (arxiv.org).
Ollama supports Qwen / Shared entities / Shared topic / Earlier coverage / Tension
Linked by a graph relationship (Ollama supports Qwen); both cover Qwen, SWE; overlapping topics (code, cost).
OpenCode uses Qwen / Shared entities / Shared topic / Earlier coverage
Linked by a graph relationship (OpenCode uses Qwen); both cover OpenCode, Qwen; overlapping topics (against, agent, code).
Linked by a graph relationship (OpenCode uses Qwen); both cover OpenCode, Qwen; overlapping topics (agent, code, cost).
Qwen benchmarked against Claude / Shared entity: SWE / Same source domain / Shared topic / Earlier coverage
Linked by a graph relationship (Qwen benchmarked against Claude); both cover SWE; reported by the same outlet (arxiv.org).
OpenCode uses Qwen / Shared entity: Opencode / Same source domain / Shared topic / Earlier coverage
Linked by a graph relationship (OpenCode uses Qwen); both cover Opencode; reported by the same outlet (arxiv.org).
Alibaba released Qwen / Shared entities / Shared topic / Earlier coverage
Linked by a graph relationship (Alibaba released Qwen); both cover Qwen, SWE; overlapping topics (against, agent).
OpenCode uses Qwen / Shared entities / Shared topic / Earlier coverage
Linked by a graph relationship (OpenCode uses Qwen); both cover OpenCode, Qwen; overlapping topics (agent, code).