SWE-rebench

mentions 2 type Organization feed RSS

// recent coverage 2 mentions

17:09

2026-07-31

sourcefeed.dev

artificial-intelligence

Frontier Coding Models Are Two Points Apart. Costs Aren't

SWE-rebench's latest leaderboard shows frontier coding models within 2.2 points of each other on resolved rate, with Fable 5 (Anthropic) at 64.5%, Grok 4.5 (xAI) at 63.8%, Opus 5 (Anthropic) at 63.4%,…

22:32

2026-06-17

transluce.org

ai-agents

A framework for verifiable analysis of AI behavior

Transluce Labs released Docent, a framework for verifiable analysis of AI behavior that allows humans to inspect and audit every step of an AI agent's analysis. The framework uses analysis plans with …

// co-occurs with top 8 entities

OpenAI 2 Transluce Labs 1 Docent 1 Claude Code 1 GitHub 1 GPT-5.5 1 Nebius 1 Anthropic 1

// topics top 6 topics

ai tools 2 ai research 2 ai agents 1 ai safety 1 ai ethics 1 artificial intelligence 1