Anthropic's safer Mythos: 2x the price, still a risky gamble
anthropic released a 'safe' version of its mythos model with doubled pricing and strict user limitations. this strategic shift towards controlled access reflects the lab's attempt to balance safety with commercial deployment.
anthropic released a 'safe' version of mythos, a model it previously refused to release, featuring strict guardrails and a price tag double that of its last flagship. the firm claims the new user limitations make the system safe for public use, though critics suspect the initial danger warnings were a marketing tactic to build hype. selective access, rather than full open-sourcing, is emerging as the dominant strategy for frontier model labs balancing capability with risk. this release signals a trend where safety is directly tied to price and access control.