What is Fable 5.2?
Fable 5.2 is Anthropic's new model, leading the Remote Labor Index at 41.3%. See how it compares with Fable 5.1, Opus 5.5, and GPT 6 Astra.
Updated ·
Fable 5.2: Anthropic's Model That Leads the Remote Labor Index
Announcement date:
Fable 5.2 is a new Anthropic model that tops the Remote Labor Index (RLI) from Scale Labs and the Center for AI Safety with a score of 41.3 percent, meaning it completes more than 41 percent of real freelance projects at professional quality. That is more than twice the rate of its direct predecessor, Fable 5.1, which sits at just under 18 percent. The result marks a sharp jump from a benchmark where the leading system managed only 2.5 percent when it launched about a year earlier.

What Fable 5.2 is and who makes it
Fable 5.2 is a model from Anthropic, released as the successor to Fable 5.1 in the same model family. It is positioned around agentic work: the kind of multi-step, tool-using tasks that resemble paid freelance projects rather than short chat prompts.
The headline claim attached to the release is benchmark performance. On the Remote Labor Index, Fable 5.2 completes more than 41 percent of real freelance projects at professional quality, the highest score reported on the current leaderboard.
- Developer: Anthropic
- Family: Fable, succeeding Fable 5.1
- Headline result: 41.3 percent on the Remote Labor Index
- Focus: real freelance-style project work rather than short prompts
How Fable 5.2 compares on the Remote Labor Index
On the RLI from Scale Labs and the Center for AI Safety, Fable 5.2 leads with 41.3 percent. Its direct predecessor, Fable 5.1, trails far behind at just under 18 percent, so Fable 5.2 completes more than twice as many jobs.
The next spots on the current leaderboard go to Anthropic's Opus 5.5 at 21.25 percent and OpenAI's GPT 6 Astra at 20.83 percent. Google's strongest model in the test, Gemini Argon 4, lands in fifth place at 17.08 percent, while older models such as GPT-5.2 and Gemini 3 Pro sit below three percent.
- Fable 5.2 (Anthropic): 41.3 percent
- Opus 5.5 (Anthropic): 21.25 percent
- GPT 6 Astra (OpenAI): 20.83 percent
- Fable 5.1 (Anthropic): just under 18 percent
- Gemini Argon 4 (Google): 17.08 percent, fifth place
- GPT-5.2 and Gemini 3 Pro: below three percent
Where Fable 5.2 improves most, and where it still struggles
Scale Labs said on X that Fable 5.2 makes its biggest gains in editing audio and music, designing architecture and interior spaces, and creating 3D models. Nearly 50 of its successful completions were jobs that no AI agent in the test had solved before, which suggests the gains are not only incremental improvements on familiar task types.
The same source notes clear remaining weaknesses. Scale Labs says the AI can still barely handle other tasks, especially industrial product design and animations. Those areas are worth treating as unreliable for now rather than as solved capabilities.
- Strongest reported gains: audio and music editing, architecture and interior design, 3D modeling
- Nearly 50 successful completions were tasks no agent in the test had solved before
- Still weak: industrial product design and animations
What the Remote Labor Index measures
The RLI is a benchmark from Scale Labs and the Center for AI Safety built from real paid work rather than synthetic exercises. Its dataset includes 240 completed projects from 358 verified professionals on the Upwork platform, together worth about $144,000.
The scale of individual tasks matters for interpreting scores. A single project costs an average of $632.60 and takes a human freelancer an average of just under 29 hours, so a percentage point on this benchmark represents substantial, multi-hour work rather than a quick question.
The benchmark's history also frames the Fable 5.2 result. When the RLI launched about a year earlier, the top system at the time, Manus, managed just 2.5 percent of tasks. In July, the older Fable 5 was still in front at about 16 percent.
- Built from 240 completed Upwork projects by 358 verified professionals
- Total value of the included jobs: about $144,000
- Average project: $632.60 and just under 29 human hours
- Launch-era leader Manus: 2.5 percent
- July leader Fable 5: about 16 percent
Getting started and what is still unknown
This profile is based on benchmark reporting and Scale Labs commentary. The supplied material does not document Fable 5.2's pricing, context window, API identifiers, rate limits, or general availability, so those details should be confirmed on Anthropic's official channels before planning around them.
A practical way to evaluate the model is to test it against the task types where it reportedly gained most, such as audio editing, spatial design, and 3D modeling, and to keep human review in place for industrial product design and animation work, where the benchmark still shows weak performance.
- Pricing, access tiers, and limits: not documented in the supplied material
- Suggested evaluation: trial on audio, architecture/interior, and 3D tasks
- Suggested caution: keep review on industrial product design and animation
Questions
What is Fable 5.2 and who makes it?
Fable 5.2 is a model from Anthropic and the successor to Fable 5.1. It leads the Remote Labor Index with 41.3 percent, completing more than 41 percent of real freelance projects at professional quality.
How does Fable 5.2 compare with Fable 5.1, Opus 5.5, and GPT 6 Astra on the RLI?
Fable 5.2 scores 41.3 percent, more than twice Fable 5.1 at just under 18 percent. Anthropic's Opus 5.5 follows at 21.25 percent and OpenAI's GPT 6 Astra at 20.83 percent.
Which tasks does Fable 5.2 handle best, and where is it still weak?
Scale Labs reports the biggest gains in editing audio and music, designing architecture and interior spaces, and creating 3D models. It says the model can still barely handle industrial product design and animations.
What is the Remote Labor Index?
The RLI is a benchmark from Scale Labs and the Center for AI Safety built from 240 completed Upwork projects by 358 verified professionals, worth about $144,000 in total. A single project averages $632.60 and just under 29 human hours.
Original sources
Tool Info
Has this tool changed? Help us keep it up to date
Fable 5.2 Reviews
Loading reviews…