STDStandardsModels

SWE-bench Verified

2 mentions across all digests

SWE-bench Verified is a benchmark that evaluates AI coding agents on their ability to resolve real GitHub issues in open-source repositories; OpenAI announced they stopped using it as a primary evaluation benchmark, signaling potential saturation concerns.

/// Stats

First Seen2026-03-24

Last Seen2026-03-24

Total Mentions2

Subject Mentions1

Last 7 Days0

Sources1

Peak Relevance4/5

Active Predictions0

/// Recent Stories

2026-03-21HIGH

Why we no longer evaluate SWE-bench Verified

2026-03-21HIGH

Building more with GPT-5.1-Codex-Max

/// Connected Entities

ORGOpenAI

2 shared