X
Publisher · verified by editors
X: @lmsysorg
AI news source. Articles are auto-selected and adapted by Hamidun News editors.
Latest publications
AutonomousX: @lmsysorg
ApprenticeBench: agents outperform human testers but cost 2.5 times as much
A new benchmark tests whether AI agents can settle into a job like new hires: learning internal rules, tools and unwritten knowledge. The result: better accuracy than humans, at a higher cost.
HardwareX: @lmsysorg
SGLang adds support for gpt-oss-120b on Hopper and Blackwell GPUs
OpenAI's open model gpt-oss-120b now runs in SGLang on two generations of Nvidia GPUs. What that means if you host your own LLMs and track the cost per token.