Gestalt AI中文

Gestalt AI model overview & benchmark scores

Last updated:

Gestalt is an AI design model built for real-world websites and application interfaces. In the Frontier Design Benchmark by Contra Labs and BenchFlow, it ranked #1 in double-blind evaluations by professional designers.

Top 8 leaderboard comparison

The November 18, 2026 research report recorded Gestalt at 1,928 Elo. The table above reflects the Top 8 live leaderboard as checked on November 21, 2026. Win rate reflects overall blind head-to-head comparisons.

Unlike synthetic coding evaluations, Frontier Design Benchmark tests end-to-end commercial interface design across layout hierarchy, visual craft, and brand alignment.

RankEntryElo (Nov 21, 2026)GapWin rate95% CI
#1Gestalt (stealth)1,931Top90%86–92%
#2Human designers1,739-19274%71–78%
#3Claude Fable 5.21,713-21873%69–76%
#4Claude Opus 5.51,691-24073%—
#5GPT-6.1 Sol1,668-26370%66–74%
#6Claude Sonnet 5.51,667-26469%65–73%
#7Kimi K3.1 Pro1,663-26871%66–74%
#8GPT-6 Astra1,638-29366%61–70%

Design quality: rubric score vs blind preference

On the benchmark’s 35-dimension Design Critique Rubric (which correlates with expert preference at r = 0.94), Claude Fable 5.2 and Claude Opus 5.5 score 84, Gestalt scores 83, and human reference designs score 74.

Even with similar rule-level rubric totals, Gestalt leads Claude Fable 5.2 by +218 Elo in blind designer voting and sits close to human references on the Slop Detector (0–100)—pointing to stronger whole-page composition and fewer generic AI layout clichés.

Where to view outputs and current release note

Side-by-side generated websites from all evaluated entries can be inspected directly in the Contra Labs Design Gallery (contralabs.com/frontierdesign/gallery).

Gestalt currently participates in the benchmark as a stealth entry, with no public API or release date announced yet.

FAQ

How far ahead is Gestalt AI on the Frontier Design Benchmark?

At 1,931 Elo (90% win rate, 95% CI 86–92%), Gestalt leads human reference designs by +192 Elo, Claude Fable 5.2 by +218 Elo, and GPT-6.1 Sol by +263 Elo.

Why does Gestalt beat models with a similar rubric score?

While top models score 83–84 on formal checklist criteria, Gestalt pairs high craft with much lower AI-slop scores and stronger holistic visual hierarchy, winning 90% of blind head-to-head comparisons.

Where can I see websites designed by Gestalt?

Contra Labs publishes benchmark outputs side by side in the official Frontier Design Benchmark gallery (contralabs.com/frontierdesign/gallery).

Gestalt AI

Model overview