‹ PublicAI Index
The LLM benchmark aggregator.
Claude Fable 5.2
Anthropic
Strongest in Agents (#1 of 298, on 3 of its 6 boards), weakest in Legal (#95 of 173). Above par in 16 of 17 scopes. Among the models it meets almost everywhere, it finishes behind Claude Opus 5.5 and Claude Fable 5 and ahead of GPT-6 Astra and Claude Fable 5.1.
§ 1 · Profile
What it is good at
Bars run from 50 — the average of the models each source lists — so right of the line is above par. The middle column is the gap to whoever leads that scope.
Agents66.4leads#1/2983/6
Long-horizon work74leads#1/311/1
Web research62.7−1.2#2/351/2
Knowledge work70.2−0.5#3/2082/2
Computer use60.4−7.7#15/681/2
Core abilities64.5−0.8#3/1992/4
General intelligence69.4leads#1/1782/3
Data analysis60.4−4.8#6/311/1
Reasoning63.3−2.1#6/1954/5
Reasoning65.5−3.5#6/1473/3
Mathematics61.2−2.6#8/1521/2
Professional56.3−6.1#22/1942/2
Document design68.7leads#1/311/1
Biology research70.7leads#1/311/1
Legal49.5−16.4#95/1731/1
Coding57.4−8.8#23/1792/6
Agentic coding57.9−9#21/1732/5
§ 2 · Head to head
What it beats, and what beats it
The same models turn up scope after scope. Counted once: where both were placed, who finished higher. Bars run right for Claude Fable 5.2, left for the other.
§ 3 · Sources
Where the numbers come from
7 publications, 18 figures. Every one links to the page it was read from.
Artificial Analysis Intelligence Index 65
GDPval-AA 1831
AA-Briefcase 1785
ARC-AGI-2 92.5%
LiveBench 84.9LiveBench · Agentic 82.9LiveBench · Reasoning 88.8LiveBench · Coding 79.2LiveBench · Browser Use 86.9LiveBench · Data Analysis 81.5LiveBench · Research 94.5LiveBench · Design 85.1
SimpleBench 94.9%
Vals · Legal Research Bench 46.15%Vals · Harvey Legal Agent Benchmark 5%Vals · BioMysteryBench 83.7%Vals · Vibe Code Bench 93.77% (OpenHands)Vals · ProofBench 100%
Badge
[](https://publicai.io/model-index/m/claude-fable-5-2)