Test: Meta's Muse Wonder nearly matches Claude Fable 5.1 on agentic tasks

Meta launched Muse Wonder on October 12: the standard model is available via Model API and Muse Code at $4/$20 per million tokens, while Muse Wonder Full is limited to top Muse Code plans. In head-to-head tests on bug fixing, incident investigation and long-session memory, Muse Wonder matched Claude Fable 5.1 but used about twice as many tokens and cost more.

• Muse Wonder costs $4 per million input and $20 per million output tokens, 60% cheaper than Fable 5.1 at $10/$50
• In the bug-fix test Muse passed 5 of 6 hidden tests; Fable 5.1 passed all 6
• Incident investigation: both scored 10/10, but Muse took 13 turns versus Fable's 6
• In the memory test Muse scored 25/25, while Fable refused 34 of 50 turns due to a cyber filter