Anthropic launches Claude Opus 5 between Sonnet and Fable

Anthropic launched Claude Opus 5 on July 24 and placed it between Sonnet and Fable in its expanding model lineup. The company said Opus approached Fable capability at roughly half Fable's price. That comparison is Anthropic's positioning and should be tested on specific tasks rather than treated as a universal performance result.

Opus 5 became the default model for Max subscribers and the strongest model available to Pro users, according to Anthropic. Plan access matters because the same person may encounter a different default depending on subscription. Entitlements and usage limits can change, so a current account check should accompany any buying guide.

The company emphasized coding, professional knowledge work, verification, iteration and an adjustable effort setting. An effort control can help users trade time and cost against depth, but it also complicates comparisons. Two results from the same named model may use different effort levels and should not be presented as equivalent runs without that setting.

Reuters and Axios reported the lower-cost positioning and Anthropic's rapid release schedule. Their accounts provide independent confirmation of the launch and market framing, not independent reproduction of Anthropic's capability tests. Claims that Opus is close to Fable need the workload and evaluation conditions attached.

The practical decision is whether Opus saves enough supervision or improves enough output to justify its cost over Sonnet, and whether Fable adds value beyond it. A useful comparison should include token price, effort, latency, retries and review time. Model names alone cannot answer which tier is economical for a given team.