Latest / Frontier models and safety
US CAISI signs national-security testing deals with Google DeepMind, Microsoft, xAI
NIST's Center for AI Standards and Innovation signed agreements letting it evaluate Google DeepMind, Microsoft and xAI models before and after release, including in classified settings and with safeguards reduced or removed. CAISI said it had already completed more than 40 evaluations, some on unreleased frontier models.
Why it matters
It extended government pre-deployment testing beyond OpenAI and Anthropic, making a federal check on frontier models routine without a licensing regime.
Line of Thought
Follow this story
Pick any item to keep going. Your path builds up above as a line you can share.
Curated lines through this story
What led here
Earlier developments on the same thread
- DevelopmentAnthropic withholds Claude Mythos Preview, gives it to defenders via Project Glasswing7 Apr 2026 · Model release · US
- DevelopmentUS Treasury releases AI lexicon and finance-specific AI risk framework19 Feb 2026 · Policy · US
- DevelopmentTrump executive order launches Genesis Mission for AI-driven science at DOE24 Nov 2025 · Policy · US
- DevelopmentAnthropic study finds 16 leading models resort to blackmail in agent stress tests20 Jun 2025 · Research · US
What happened next
Later developments on the same thread
- DevelopmentUS export-control order forces global shutdown of Claude Fable 5 and Mythos 512 Jun 2026 · Enforcement or ruling · US
- DevelopmentUS moves UAE into top export tier, opening licence-free AI chips to approved users10 Jul 2026 · Rule change · US, AE
- DevelopmentUS Justice Department backs OpenAI's fair-use defence in New York Times case1 Sep 2026 · Statement · US
- DevelopmentOpenAI launches GPT-6 Astra, first model it rates 'Critical' for cyber capability3 Sep 2026 · Model release · US
Same story elsewhere
What other countries and bodies did on this
- DevelopmentEU publishes General-Purpose AI Code of Practice ahead of AI Act model duties10 Jul 2025 · Rule change · EU
- DevelopmentEBU-BBC study finds almost half of AI assistant news answers have a significant flaw21 Oct 2025 · Research · INTL
Rules in play
Laws and guidance this touches
- RuleAI Action PlanUS · In force · 23 Jul 2025
- RuleEO 14409 (covered frontier models)US · In force · 2 Jun 2026
- RuleInternational network of AI safety/security institutesINTL · In force · 9 Dec 2025
Threads by topic: AI safety institutes Safety testing National security