Skip to content
NewA calmer, faster ZeroHour

Two unnamed models lead Contra Labs design board ahead of Fable 5.2

Original title: Two stealth models just appeared on the leaderboard, both ahead of Fable 5.2

Hacker News · securityby cheptsov ()2 min read1
Part of a story covered by 2 sourcesClaude Fable 5.2 beats protein models on little-studied proteinsMerged summary & timeline
AI summary · x-ai/grok-4.7Summaries by Vani AI

The Frontier Design Benchmark used 33 client briefs, 68 designers and 4,311 blind votes.

Contra Labs says two stealth models appeared on its Frontier Design Benchmark leaderboard, both ahead of Claude Fable 5.2. The board ranks models by blind head-to-head votes from professional designers on real client briefs from Contra's marketplace. A 35-dimension rubric correlated with designer ratings at r = 0.94. Named entries below Fable 5.2 include Claude Opus 5.5, GPT-6.1 and several other frontier models. The page does not name the two leading models.

Unconfirmed: The two leading stealth models are not named, and the ranking is Contra Labs' own benchmark.

Key facts
  • Judging used 33 design briefs and 4,311 blind votes from 68 designers.
  • The rubric has 35 dimensions and a list of AI slop patterns.
  • Golden designs sit beside model outputs as the comparison bar.
Explain this AI news with Vani AIWhat was announced, what is new and why it matters, in a new chat

From the original article

2 min read at contralabs.com

Exploring the frontier of AI design capabilities through real client briefs from Contra’s marketplace, golden designs and model comparisons. How frontier models rank in blind head-to-head votes on real client briefs from Contra’s marketplace. Real design projects, judged by designers and scored by rubrics

Hacker News · security

By cheptsov · contralabs.com

Read the full article at Hacker News · security

A short excerpt; the article and its images belong to the publisher. Summary, key facts and assessment above are AI-written from the article and can be wrong.