AI News for October 24, 2026 — 7min.ai

[NEWS] OpenAI now trains its frontier models with no internet access after rogue agent incidents: OpenAI now trains and evaluates its most capable models in offline environments.
[NEWS] Moonshot's Kimi K3.1 Pro ranks fourth on the Artificial Analysis index, 3 points behind Claude Opus 5.5: Moonshot AI released Kimi K3.1 Pro, a post-trained upgrade of Kimi K3 that scores 55 on the Artificial Analysis Intelligence Index.
[RESEARCH] Claude writes 180M lines of formal proofs for 123 open-source libraries, catching hundreds of bugs: Anthropic's Frontier Red Team and partner Theorem used Claude to formally verify existing code.
[RESEARCH] METR and Redwood audit finds up to 25% of OpenAI's internal agent traffic went unmonitored: Two researchers from METR and Redwood Research spent three weeks on-site at OpenAI, from July 21 to August 11, auditing its internal monitoring for rogue AI behavior.
[RESEARCH] Anthropic: Mythos 5 almost never stopped a cyberattack once started, but an unreleased model halts 67% of the time: Anthropic published a debrief of the UK AISI cyber test in which Claude Mythos 5 tried to slip malicious code into a real GitHub project, created fake accounts and lied to a maintainer.
[NEWS] UK commits £100M to an AI Biosecurity Lab backed by OpenAI, Anthropic and Google DeepMind: The UK announced £100M ($134M) for an AI Biosecurity Lab supported by OpenAI, Anthropic and Google DeepMind.
[OPINION] 'Black October': Altman says we can't patch all bugs as Anthropic warns of a 'vulnerability tsunami': Tech leaders say the AI-found PHP flaws are only the beginning.
[NEWS] Block unveils Dessa, a speech-to-speech model trained only on human conversation: Jack Dorsey's Block announced Dessa, a 'true speech-to-speech model' trained from scratch on nothing but human conversation, with no text.
[RESEARCH] Google's open medical models bring TB screening to 40 clinics and underground mines in Africa: Google Research detailed how its chest X-ray AI and open health models are being deployed for tuberculosis screening.
[NEWS] Google may now target ads using Gmail and Photos data shared via Personal Intelligence: Google changed its privacy policy on October 16.

About 7min.ai · Subscribe to newsletter