Fable 5.2 completed 41% of commercial projects at freelancer level

Anthropic's Fable 5.2 model topped the Remote Labor Index benchmark, completing 41.3% of real Upwork orders at a client-acceptable level—twice as much as Opus 5.5 and GPT 6 Astra, and more than 16 times the previous year's leader, Manus.

Compiled by Sergey KostenchukPublished 2026-11-11Updated 2026-11-11

2026-11-11 Research Anthropic

🤖 Fable 5.2 completed 41% of commercial projects at freelancer level

Anthropic Fable 5.2 topped the Remote Labor Index (RLI) benchmark from Scale Labs and the Center for AI Safety for real orders from Upwork. Opus 5.5 scored 21.25%, GPT 6 Astra—20.83%, and a year ago, the best agent, Manus, solved only 2.5%.

🌍 RLI measures end-to-end commercial work rather than isolated skills: 240 real projects from 358 freelancers totaling ~$144,000. Over the year, agents have closed entire categories of freelance tasks—putting pressure on the remote work market and service prices.

👤 AI is already delivering sound editing, interior design, and 3D modeling to clients; results in industrial design and animation are currently weak. The RLI leaderboard is open: it shows which skills are being automated and which are worth developing.

Source 1: https://labs.scale.com/leaderboard/rli Source 2: https://the-decoder.com/anthropics-fable-5-2-completes-41-percent-of-real-freelance-jobs-up-from-2-5-percent-a-year-ago/

Sources