Anthropic confirms Claude Mythos-class models launch for public
▼ Summary
– Anthropic plans to release Mythos-class models to the general public in the coming weeks after delaying the rollout due to security risks.
– Mythos was announced in April as a restricted model available only to select companies and security researchers.
– Anthropic has reportedly developed strong guardrails to prevent misuse of the Mythos model, which is believed to be more powerful than Opus 4.8.
– The Mythos model shows major improvements in code reasoning and autonomy compared to Claude’s current flagship model, Opus 4.8.
– A “Mythos-preview” model briefly appeared for some users on Claude Code before being taken offline.
Anthropic has officially confirmed plans to release its Mythos-class models to the general public, following a delay driven by concerns over potential risks to both public and private software systems.
Initially unveiled in April, Mythos was launched as a restricted model available only to a select group of organizations, including security researchers. At the time, Anthropic highlighted significant security vulnerabilities associated with the model and chose to hold back a wider release.
“The advantage will belong to the side that can get the most out of these tools,” the company stated in April when announcing the model. “In the short term, this could be attackers, if frontier labs aren’t careful about how they release these models. In the long term, we expect it will be defenders who will more efficiently direct resources and use these models to fix bugs before new code ever ships.”
AI developers typically delay the rollout of highly capable models until robust guardrails are in place to prevent misuse. Anthropic now appears to have developed those safeguards for Mythos, which is believed to be significantly more powerful than Opus 4.8 and other currently available models.
In a recent blog post, the company confirmed it intends to release Mythos-class models to the public within the coming weeks, though it has not set a specific date. “We’re making swift progress on developing these safeguards and expect to be able to bring Mythos-class models to all our customers in the coming weeks,” Anthropic wrote.
Currently, a small number of organizations are already using a Claude Mythos preview for cybersecurity tasks, but it remains unclear whether the same version will be offered to the broader public. According to Anthropic, the Mythos model demonstrates major improvements in code reasoning and autonomy, surpassing the capabilities of Claude’s current flagship, Opus 4.8.
Interestingly, the “Mythos-preview” model briefly appeared for some users on Claude Code before being taken offline, hinting at the system’s readiness for a wider launch.
(Source: BleepingComputer)
