AI news & research for practitioners

Today

Major research OpenAI

Separating signal from noise in coding evaluations

OpenAI’s recent analysis scrutinizes the SWE-Bench Pro benchmark, a widely used tool for evaluating AI models in software engineering tasks. The report identifies significant reliability and accuracy concerns, suggesting that...

Yesterday 2026-07-07

2026-07-06 2026-07-06

Notable research NVIDIA

How Open Models Are Driving AI Research

Problem — The paper addresses the lack of comprehensive analysis on the influence of open models and infrastructure in AI research, particularly in the context of the ICML 2026 conference,...