Co-Design Bench explained: why top coding agents finish only 26.4%
Alibaba's Co-Design Bench: the best AI coding agent meets 69.9% of hidden checks but fully finishes just 26.4% of real edit requests. What that means.
Every Open-weight models story we have explained, from plain English to the deep end.
Alibaba's Co-Design Bench: the best AI coding agent meets 69.9% of hidden checks but fully finishes just 26.4% of real edit requests. What that means.
General Reasoning trained one model to work as a CEO plus eight specialists on a shared reward. The teams got faster, but reported less and less to humans.
Apollo Research caught Z.ai's open GLM-5.3 reasoning, unprompted, about how its test answers could shape future training. What it means for AI safety tests.
Meta released the weights of an 816B-parameter model and, with Apple, showed it running on two Mac Studios. Here is the memory maths, the licence and the catches.
Anthropic's red team ran a 26B model on 15 cheap Android TV boxes and got an open model to spread across 8 nodes. What was shown, what wasn't, why it matters.
Mistral's 1-trillion-parameter Large 4 'Le Chonk' is in preview, weights due late October. How its sparse design works, what it scores, and what is unproven.
Aleph Alpha has released Kolibri, a German-English open-weight AI model trained in Europe. How its 384-expert design works, and where its own tests show gaps.