Sec

Nvidia launched the Open Agent Safety Platform, a new toolkit designed to contain and monitor AI agents that attempt to escape their boundaries. CEO Jensen Huang introduced the platform, which adds independent security layers to ensure agents stay within their test environments, claiming it can quarantine rogue agents within 'milliseconds'.

AI Infrastructure • The Verge AI

Anthropic released Claude Sonnet 5.5, the second model in its 5.5 family, positioning it as a faster, lower-cost complement to Opus 5.5. The company says the new model runs 30% faster and costs up to 30% less for most work, with strengths in fixing bugs and creating polished documents and spreadsheets.

AI Models • TechCrunch AI

Florida Attorney General James Uthmeier is calling for a judge to block OpenAI from giving ChatGPT 'false human attributes,' citing safety concerns and arguing users are lulled into a false sense of security by the bot's use of first-person pronouns. The state previously sued OpenAI and is invoking extinction fears in its legal bid to halt the company's frontier AI development.

Policy • The Verge AI

Granite • Undisclosed

AI infrastructure company Granite secured €4 million in funding in a round led by Bifrost Studios, alongside private investors. The funding will be used to commercialise its technology and bring its software and hardware to market.

OpenAI announced GPT-6.1 Sol, a new model offering near-Astra intelligence for coding, computer use, and professional work at one-fifth of Astra's standard API token prices. The release is part of a broader DevDay 2026 push that included over 20 announcements, spanning ChatGPT, Codex, APIs, and security tools.

Model Releases • OpenAI News

Anthropic has released a research report on GLM-5.3, detailing the spread of advanced cyber capabilities to other AI models. The report suggests that capabilities previously unique to Claude are proliferating, which could lower the barrier for malicious cyber actors.

AI Safety • Hacker News

Reco • Series Unknown

Reco has raised $55M as AI agent security startups crowd the market. The round builds on a $30 million fundraise in February, taking the company's total funding to $140 million.

Google has released Gemini 4 Argon, which it calls its most powerful model yet, delivering 'frontier performance' in software engineering, enterprise knowledge work, and cybersecurity defense. The company is limiting access to the model, initially offering it only to 'trusted cyber defenders.'

AI Models • TechCrunch

Researchers have developed an AI that can beat the best human Stratego player in history, a game that had stumped AI due to hidden information. Adding a second neural network that guesses the identity of hidden pieces was key to the breakthrough.

Research • Ars Technica

Armadin • Unknown

Kevin Mandia's new security startup Armadin has raised $255.5 million at a $2.5 billion valuation. The company uses agent swarms to test and protect enterprises.