Editorial illustration for Anthropic's Fable 5.2 completes 41% of freelance jobs, up from 2.5% a year ago

Anthropic's Fable 5.2 Completes 41% of Freelance Jobs

By AI Daily Post Edited by Brian Petersen, Editor-in-Chief

November 10, 2026 • 4 min read

Anthropic's latest AI model, Fable 5.2, just completed 41.3 percent of real freelance jobs on the Remote Labor Index, a benchmark run by Scale Labs and the Center for AI Safety that tracks whether AI agents can finish actual paid work to a standard clients would accept. A year ago, the top performer on that same test, a system called Manus, managed just 2.5 percent. Fable 5.2 is now clearing more than sixteen times that bar.

The climb didn't happen in a straight line. Fable 5, an earlier model in the same lineup, led the pack in July at roughly 16 percent, already a fourfold jump from where things stood eight months before. Then came Fable 5.1, which only nudged that number to just under 18 percent. Fable 5.2 arrived four months after the July mark and more than doubled Fable 5.1's score in one release.

What that pace of improvement says about how close AI agents are to doing freelance-grade work at scale, and what it means for the people currently doing that work for a living, is the question the Remote Labor Index was built to answer.

On the Remote Labor Index (RLI) from Scale Labs and the Center for AI Safety (CAIS), Anthropic's new model Fable 5.2 takes the lead with 41.3 percent. The test measures how often AI agents complete real freelance jobs well enough for a paying client to accept the work.

Anthropic's Fable 5.2 completes 41 percent of real freelance jobs, up from 2.5 percent a year ago - THE DECODER

Why this matters

A jump from 2.5 percent to 41.3 percent in a year is the kind of curve that should make anyone pricing freelance work pay attention. The Remote Labor Index isn't a synthetic benchmark; it's built from actual client jobs, which means Fable 5.2's gains in 3D modeling, audio editing, and architectural design translate directly into categories where human freelancers currently earn a living. For founders building on Anthropic's stack, the signal is that agents are starting to clear the bar on sustained, multi-step creative work, not just narrow coding tasks. That's a different kind of capability than passing a trivia test.

We'd still push back on reading 41 percent as "nearly half of freelance work is automated." Scale Labs and CAIS haven't published enough detail here on failure modes or how "professional quality" was judged, and benchmark leaders have a habit of overstating real-world transfer. But the trajectory, three model generations to go from token completion rates to over 40 percent, is the number worth tracking, not the snapshot.

Common Questions Answered

What is the Remote Labor Index and how does it measure AI agent performance?

The Remote Labor Index (RLI) is a benchmark developed by Scale Labs and the Center for AI Safety that tracks whether AI agents can complete actual paid freelance work to a standard that paying clients would accept. Unlike synthetic benchmarks, the RLI is built from real client jobs, providing a practical measure of AI capability in actual work scenarios.

How much has Anthropic's Fable 5.2 improved compared to previous AI models on the Remote Labor Index?

Fable 5.2 achieved a completion rate of 41.3 percent on the Remote Labor Index, compared to just 2.5 percent for Manus, the top performer a year ago. This represents a more than sixteen-fold improvement in AI agents' ability to complete real freelance jobs over the course of a year.

Which specific freelance job categories has Fable 5.2 shown improvements in?

Fable 5.2 has demonstrated significant gains in 3D modeling, audio editing, and architectural design according to the article. These improvements in specialized freelance categories directly impact areas where human freelancers currently earn their living.

Why should freelance workers be concerned about Fable 5.2's performance on the Remote Labor Index?

The jump from 2.5 percent to 41.3 percent completion rate in just one year represents a significant acceleration in AI agents' ability to perform real paid work. Since the Remote Labor Index measures actual client jobs rather than synthetic tests, Fable 5.2's gains directly translate to categories where human freelancers currently earn income, potentially disrupting the freelance labor market.

Ship an AI product this weekend — no engineers required.

Structured, in-depth lessons on the exact no-code tools — not scattered tutorials.

The exact platforms, taught in depth
Build real, working projects
Our honest review + a reader discount

Read the review →