Anthropic restores two models and adds controls for enterprise teams
Reviewed by Helen Jones
Anthropic restored access to Claude Fable 5 and Mythos 5 on July 1. The same day it released Claude Sonnet 5. Enterprise plan administrators can now set which models teams may use and at what effort level. Mythos 5 remains restricted to vetted US partners under Project Glasswing.
Before this update, founders either accepted whatever model Anthropic made available or switched providers when a model disappeared. The suspension of Mythos-class models in June showed how quickly an entire capability tier could vanish. Now model choice becomes an explicit operating parameter. A solo founder can route routine support replies or first-draft copy to the cheapest model that meets quality thresholds while reserving higher-cost models for offer pages or customer-facing sequences that directly affect conversion.
Analysis
Stop treating model selection as a background default. Map every recurring prompt chain to the lowest-cost model that still passes a human quality check, then lock that assignment in your workflow so the next model release does not silently inflate your token bill.
Pulse published by Collab365 Spaces, reviewed by Helen Jones on . Cite as "Anthropic restores two models and adds controls for enterprise teams", Collab365 Spaces. 2 sources referenced.