[AINews] Agents disprove Polysemanticity; greatly improve Interpretability & Alignment
Meta released Muse Spark Open, an 816B MoE that leads US open-weight models on Artificial Analysis.
Meta released Muse Spark Open, an 816B-total / 42B-active mixture-of-experts model with a 1M-token context and text, image, video, and audio input. Artificial Analysis scores it 39, first among US open-weight models but behind GLM-5.3 (45) and Kimi K3 (44); launch pricing is $1.25/$4.25 per million input/output tokens. The same roundup covers a Goodfire interpretability essay and launches of NVIDIA Nemotron 3.5 Super VL and Microsoft’s roughly 3T-parameter MAI-Base-2.
78