Anthropic chose not to release its first model, Claude 1, before ChatGPT despite seeing its power. They worried it would trigger a dangerous "arms race" and decided the commercial cost of waiting was worth the potential safety benefit for the world.

Previously, labs like OpenAI would use models like GPT-4 internally long before public release. Now, the competitive landscape forces them to release new capabilities almost immediately, reducing the internal-to-external lead time from many months to just one or two.

The timing of OpenAI's next model release relative to Anthropic's upcoming 'Mythos' signals its internal competitive assessment. Releasing before Mythos would be a preemptive strike, suggesting a belief that its model can't win a head-to-head comparison. Waiting until after would signal confidence in its superiority.

The US government's intervention with Anthropic's Fable 5 model signals a new era where AI labs will hold back their most capable systems from public release. This creates a consolidation of power, with only the labs and their chosen partners having access to true frontier capabilities.

Delaying public model releases isn't a real solution for pacing AI development. The critical risk lies in a lab using its most advanced, unreleased models to accelerate internal R&D, potentially leading to a private "singularity." Meaningful pacing must limit the resources dedicated to this internal recursive loop.

Commentary from early testers of GPT 5.6 revealed it had been in testing for months, meaning its training was complete before competitors' latest models were even announced. This suggests major labs like OpenAI have already developed their true next-gen models and are strategically timing their public rollouts.

Anthropic's decision to gate its Mythos model, framed as a safety precaution, also creates powerful marketing hype, drives enterprise adoption of its native tools, and makes it harder for competitors to create imitator models.

Anthropic limited its powerful Mythos model, which finds zero-day exploits, to critical infrastructure partners. While framed as a safety measure, this go-to-market strategy also creates hype, justifies premium pricing, and prevents distillation by competitors, solidifying its brand as a responsible AI leader.

Chinese AI labs are following a playbook perfected by OpenAI. They initially release open-source models to attract developers and accelerate learning. Once they approach the performance of frontier models, they switch to a closed-source strategy to monetize and capture the value.