llmrun
Coding

Frontier Design Benchmark Leaderboard

Contra Labs and BenchFlow evaluate website design using real client briefs and blind comparisons by professional designers. The leaderboard reports Elo ratings within this designer panel; these are not percentages and do not contribute to llmrun Score. Gestalt is an anonymous stealth entrant whose developer has not been disclosed.

Source: Contra Labs × BenchFlow3 open models ranked+13 proprietary

All models ranked on Frontier Design Benchmark

Proprietary / closed models are shown dimmed — you can't run them locally, but they show where the open field stands.

#ModelScoreVRAM @ Q4
1Gestalt (stealth) · proprietary
1931.0
Stealth preview
2Claude Fable 5.2 · proprietary
1713.0
API only
3Claude Opus 5.5 · proprietary
1691.0
API only
4GPT-6.1 Sol · proprietary
1668.0
API only
5Claude Sonnet 5.5 · proprietary
1667.0
API only
6Kimi K3.1 Pro · 2780.1B
1663.0
~1674 GB
7GPT-6 Astra · proprietary
1638.0
API only
8Qwen3.8 Max · proprietary
1434.0
API only
9Kimi K3 · 2779.9B
1415.0
~1674 GB
10GLM-5.3 Prime · proprietary
1403.0
API only
11Gemini 4 Carbon · proprietary
1366.0
API only
12Muse Spark 1.3 · proprietary
1335.0
API only
13Muse Wonder · proprietary
1249.0
API only
14Grok 4.7 · proprietary
1247.0
API only
15Gemini 4 Argon · proprietary
1221.0
API only
16DeepSeek V4 Pro · 1598.8B
1121.0
~960 GB

Frontier Design Benchmark: frequently asked questions

What is the best open LLM on Frontier Design Benchmark?
Kimi K3.1 Pro is the top open model on Frontier Design Benchmark, scoring 1663.0. Among all models tested — including proprietary ones — it ranks #6. The top model overall is Gestalt (stealth) at 1931.0.
Can open models match proprietary models on Frontier Design Benchmark?
Not quite on Frontier Design Benchmark: the strongest proprietary model (Gestalt (stealth)) scores 1931.0, ahead of the best open model (Kimi K3.1 Pro) at 1663.0 — but you can run the open one yourself.

Scores aggregated from Contra Labs × BenchFlow. llmrun does not run this benchmark — see the source for methodology, or the about benchmarks for what it measures.