Trends

Frontier labs gate releases behind capability thresholds and safeguard tiers

Frontier safety governance · Apr 10 – Jun 12

Organizations

AnthropicOpenAI

Updates

  1. Jun 12

    Anthropic shipped one set of Mythos-class weights twice: as Fable 5, which falls back to Opus 4.8 on cyber, bio, and distillation queries, and as Mythos 5 with cyber safeguards lifted for Glasswing partners, at $10/$50 per million tokens.

    Anthropic's Claude Fable 5 and Mythos 5

  2. May 29

    OpenAI limited its GPT-Rosalind biodefense deployment to identity-verified institutions and US and allied government partners, briefing the White House first, and published a Frontier Governance Framework aligning its Preparedness Framework with California's frontier AI act and the EU AI Act code of practice.

    OpenAI launches the Rosalind Biodefense ProgramOpenAI publishes its Frontier Governance Framework

  3. Apr 10New

    Anthropic withheld Claude Mythos Preview, which scored 93.9% on SWE-bench and autonomously found thousands of high-severity zero-days, and routed it only to Project Glasswing's defensive coalition with $100M in usage credits, the first capability hold at a frontier lab based on a documented evaluation.

    Anthropic withholds Claude Mythos Preview

Predictions

  1. Oct 2026

    At least one other frontier lab publishes a formal deployment-hold framework tied to capability thresholds, and government procurement begins requiring third-party verification of safety evaluations as a contract condition. Apr 10

    Confirmed. OpenAI limited GPT-5.6 Sol to about 20 trusted partners over its cyber capability, and Google DeepMind released Gemini 4 Argon only to vetted cyber defenders. Sources

  2. Nov 2026

    Every major frontier lab runs at least one domain-vertical deployment (biodefense, cybersecurity, industrial control, or financial systems) under a safety architecture separate from its general products, and at least one G7 government requires licensing or registration for such deployments. May 29

  3. Dec 2026

    Same-weights, different-safeguard releases with per-capability gating and government-coordinated trusted access become the default for frontier models crossing cyber or bio uplift thresholds. Jun 12