Contra Labs launches Frontier Design Benchmark; stealth model Gestalt beats human gold-standard designs
yuwen_lu_ · x · 2026-11-19
Contra Labs × BenchFlow launched the Frontier Design Benchmark, using blind head-to-head votes on real client briefs from Contra's marketplace instead of crowd votes.
Key results:
- A stealth model Gestalt tops the Elo leaderboard at 1931 with a 90% win rate — above the human golden designs (1739)
- Claude Fable 5.2 (1713), Claude Opus 5.5 (1691) and GPT-6.1 Sol (1668) round out the top five
- Kimi K3.1 Pro ranks 7th; Qwen3.8 Max 9th, while Grok 4.7, DeepSeek V4 Pro and Gemini 4 Argon land near the bottom
The site also offers design-critique and slop-detector views of each model's outputs.
More from Models
- vLLM's DSpark for Kimi K3.1 Pro: 30% longer acceptance, bounded draft KV at 1M context — AccBalanced · 2026-11-19
- Gemini 4's element naming mocked: Carbon builds on Argon, Neon on Hydrogen — teortaxesTex · 2026-11-19
- ChatGPT web translation fails badly on long Chinese text, users say free DeepSeek is better — teortaxesTex · 2026-11-19
- Rumor: Gemini 4 is a 12T-param MoE with 500B active; ChatGPT translation fumbles it — teortaxesTex · 2026-11-19
- Pangram false positives are nearly nonexistent — Bender's high-FP prediction fails a cash-bounty test — FlorianGallwitz · 2026-11-19
- Abacus.AI open-sources Smaug-Uncensored, cutting Qwen3.8-27B refusal rate from 45% to 2% — bindureddy · 2026-11-19