Model releases, research breakthroughs, and industry moves — summarized without the hype, with links to the original reporting.
AI News
Meta announced Muse Spark Open on October 15, an 816B-parameter open-weight model for coding and agents with a 1M-token context. The weights are on Hugging Face, API access runs through seven providers, and the model reportedly runs locally on two Apple Mac Studio M5 Ultra machines.
AI News
Google launched Gemini 4 Argon on October 13, built for complex multistep work in software engineering, finance, law, cybersecurity, and science. API pricing starts at $2/$10 per million input/output tokens, doubling after November 30; early testers included Deloitte and LexisNexis.
AI News
Nvidia released three fully open Nemotron 3.5 models this week on Hugging Face — Super VL (120B) for always-on agents, Bolt 74B-A9B 24GB as the balanced option, and Bolt E4B 4GB for conversational agents — shipping weights, datasets, and training recipes.
No stories match your search.