Trending storyDeveloping

Artificial Analysis ranks Claude Fable 5.2 first on AA-3D, ahead of Fable 5.1

2 articles1 sourcesince Nov 9Last article 1h ago · Nov 9, 7:41 PM ET

Overview

AISummary of 2 articles

Artificial Analysis says Claude Fable 5.2 (Max) ranks first on its AA-3D benchmark for building interactive 3D experiences, with 1271 Elo.

That is 132 points above Claude Fable 5.1 (1139), 67 points above Claude Sonnet 5.5 (1204) and Claude Opus 5.5 (1202), and ahead of GPT-6 Astra (1168) and Gemini 4 Argon (1147). The 95% confidence interval, it says, does not overlap other models.

In a separate comparison, Artificial Analysis gave Claude Fable 5.2 and 5.1 the same galaxy collision brief, tools and time limit. It says 5.2 produced a more detailed representation of two disk galaxies merging, inspired by the Antennae galaxies.

Written by AI from the articles below · updated Nov 9, 7:47 PM ET

Check the sources:

Latest developmentNov 9, 7:41 PM ET

Artificial Analysis finds Claude Fable 5.2 beats 5.1 on galaxy collision brief

Developments

2 developments

Nov 9, 7:41 PM ET · 1 article
Artificial Analysis compares Claude Fable 5.2 and 5.1
Artificial Analysis: Artificial Analysis finds Claude Fable 5.2 beats 5.1 on galaxy collision brief
Nov 9, 7:41 PM ET · 1 article
Artificial Analysis ranks Claude Fable 5.2 first on AA-3D
Artificial Analysis: Claude Fable 5.2 tops Artificial Analysis's AA-3D benchmark with 1271 Elo

Article timeline

The articles in this story. Times are ET.

Nov 9

7:41 PM
Artificial Analysis@ArtificialAnlysOfficial
Artificial Analysis finds Claude Fable 5.2 beats 5.1 on galaxy collision brief
AIArtificial Analysis compares Claude Fable 5.2, scoring 1271, with Claude Fable 5.1, scoring 1139, using the same galaxy collision brief, tools, and time limit. In its test, Claude Fable 5.2 produces a more detailed representation of two disk galaxies merging, inspired by the Antennae galaxies.
7:41 PM
Artificial Analysis@ArtificialAnlysOfficial
Claude Fable 5.2 tops Artificial Analysis's AA-3D benchmark with 1271 Elo
AIClaude Fable 5.2 (Max) ranks first on Artificial Analysis's AA-3D benchmark for building interactive 3D experiences, scoring 1271 Elo. That is 67 points ahead of Claude Sonnet 5.5 (1204) and Claude Opus 5.5 (1202), with a 95% confidence interval that does not overlap other models. It scores 132 points above Claude Fable 5.1 (1139) and leads GPT-6 Astra (1168) and Gemini 4 Argon (1147).

Heat Trend

Not enough continuous observations to show a trend yet.