Wonderful more than doubles to a $5B valuation with a $550M Series C — Insight leads again, Salesforce joins, and the bet is an AI OS for enterprise agents.
Meta Superintelligence Labs ships Muse Spark 1.3 in Muse Code and the Model API — ~20% fewer tool calls, longer-horizon agents, max reasoning still in safety testing.
HiddenLayer closes a $100M Series B led by Delta-v as enterprises fund AI runtime and agent-harness security. ARR up 10×; TechCrunch and PR Newswire detail the round.
OpenAI calls Astra its first Critical cyber model under the Preparedness Framework. Stronger safeguards after Hugging Face lessons; advanced cyber limited at launch.
Google DeepMind shipped agentic video on Gemini Flash today. It hunts frames, audio, and transcript instead of fixed 1 FPS — vendor claims up to 88% fewer tokens.
Anthropic shipped Claude Fable 5.1 and Mythos 5.1 today. Same model, two safeguard paths, 75% cheaper cache reads, and a privacy stack that keeps monitoring data on your cloud.
OpenAI is connecting authorized Epic patient charts to ChatGPT for Healthcare, plus a plugin for official public medical data. It’s read-only. It still isn’t a diagnosis.
Runway’s Solaris research model renders interactive software frame by frame. It’s a bold idea, but text, trust and accessibility are still open problems.
The Pentagon put ChatGPT Mil and Grok for Government on GenAI.mil today. IL5 workplace tools for unclassified CUI — planning, policy, logistics — not combat.
Andrew Bailey’s FSB letter to G20 finance chiefs says frontier AI’s most immediate worry for banks isn’t the hype cycle — it’s cyber risk that can move faster and hit shared tech hard.