wiki / entities / epoch-ai

Epoch AI

high confidence updated 2026-08-27 company · evaluation

Machine ingest — raw context

loading…

~… tokensappend .md to any wiki URL for this view

Epoch AI

Epoch AI publishes research on AI capabilities and benchmark interpretation. Its analysis of SWE-bench Verified emphasizes the benchmark’s concentration in small Python bug fixes, limited repository diversity, contamination risk, and sensitivity to the surrounding scaffold. [source: epoch-swe-bench-verified-analysis-2025]

agentic quality evidence — benchmark-validity evidence
error analysis and evals — evaluation design
agentic code quality — oracle-health controls

Evidence — verified primary sources

epoch-swe-bench-verified-analysis-2025

https://epoch.ai/publications/what-skills-does-swe-bench-verified-evaluate

ingested 2026-08-27

sha256:f05fee740217…

Graph context cluster: person (170 pages) betweenness: 23.2 (rank #71)

References (3)

Agentic Code Quality — Evidence Mapbenchmark-validity evidence Error Analysis and Evalsevaluation design Agentic Code Qualityoracle-health controls

Referenced by (1)

Agentic Code Quality Evidence — Entity Accountingowns the ingested SWE-bench Verified analysis.