Anthropic starts verification program for using Mythos in biology

Anthropic has opened a verification program permitting biology-related uses of Mythos, the AI model the company had restricted from life sciences applications when it released the system on July 1, according to Endpoints News. At that launch, Anthropic said Mythos was powerful enough that it could not allow the technology to be used for life sciences tasks. On Thursday, the company reversed that blanket restriction, introducing a verification process rather than lifting the limits outright.

Endpoints News reports the program as the mechanism by which biology-focused users or organizations can now gain approved access to capabilities previously walled off. The report does not specify the verification criteria in detail beyond noting the shift from prohibition to a gated approval process.

The move marks a change in how Anthropic manages the tension between the model’s stated power and its potential applications in a high-stakes scientific domain. The original restriction, announced at Mythos’s release, framed life sciences use as too risky for open access. Thursday’s announcement adjusts that stance without eliminating oversight, replacing an outright ban with a screening step.

Endpoints News is the sole source cited for this account. No further detail on the verification program’s scope, timeline, or the parties eligible to apply is given in the available material.