{"slug": "two-stealth-models-just-appeared-on-the-leaderboard-both-ahead-of-fable-5-2", "title": "Two stealth models just appeared on the leaderboard, both ahead of Fable 5.2", "summary": "Two unnamed stealth models debuted on the Frontier Design Benchmark leaderboard from Contra Labs, with the top-ranked model, Gestalt, placing ahead of the human golden reference and Claude Fable 5.2. The benchmark, built on 33 real client design briefs, 68 professional designers, 4,311 blind votes and 35 rubric dimensions, ranks Gestalt first, human golden second and Claude Fable 5.2 third, followed by Claude Opus 5.5, GPT-6.1 Sol and Claude Sonnet 5.5. Contra Labs reports its design critique rubric scores correlate with designers' ratings at r = 0.94, adding explainable quality judgment on top of pairwise preferences.", "body_md": "[Golden · Bad Move Bad Move website15 outputsClaude Fable 5.2Website](https://contralabs.com/frontierdesign/jobs/bad-move)\n\n# Deconstructing Real World Creativity.\n\nExploring the frontier of AI design capabilities through real client briefs from Contra’s marketplace, golden designs and model comparisons.\n\n## Leaderboard\n\nHow frontier models rank in blind head-to-head votes on real client briefs from Contra’s marketplace.\n\nFrontier\n\nDesign Benchmark\n\n## Elo rating from blind head-to-head votes\n\n1. Gestalt\n2. Human golden\n3. Claude Fable 5.2\n4. Claude Opus 5.5\n5. GPT-6.1 Sol\n6. Claude Sonnet 5.5\n7. Kimi K3.1 Pro\n8. GPT-6 Astra\n9. Qwen3.8 Max\n10. Kimi K3\n11. GLM-5.3 Prime\n12. Gemini 4 Carbon\n13. Muse Spark 1.3\n14. Muse Wonder\n15. Grok 4.7\n16. Gemini 4 Argon\n17. DeepSeek V4 Pro\n\ncontralabs.com/frontierdesign\n\n## Real design projects, judged by designers and scored by rubrics\n\nFrontier Design Benchmark is the first benchmark built on real world, high value, and complex design projects. It is judged by professional designers in the Contra Labs network, ensuring the quality of evaluation.\n\nIn the benchmark, we also built a design critique rubric and a list of AI slop patterns. The rubric scores are highly correlated with the designers’ ratings of the models (r = 0.94). These additional metrics provide explainable design quality judgement on top of pairwise preferences.\n\n- 33\n- design briefs\n- 68\n- professional designers\n- 4,311\n- blind votes\n- 35\n- rubric dimensions\n\n## Golden Gallery\n\nReal client briefs, each with its golden: the design set beside the models’ outputs as the bar.\n\n[View all](gallery)\n\n## Your vote ranks the models\n\nTwo model outputs for the same brief, live and side by side, without the models’ names. Pick the better one, or call a tie.\n\n[Vote on outputs](vote)", "url": "https://wpnews.pro/news/two-stealth-models-just-appeared-on-the-leaderboard-both-ahead-of-fable-5-2", "canonical_source": "https://contralabs.com/frontierdesign", "published_at": "2026-11-19 11:03:20+00:00", "updated_at": "2026-11-19 11:15:12.089562+00:00", "lang": "en", "topics": ["artificial-intelligence", "generative-ai", "ai-research", "large-language-models"], "entities": ["Contra Labs", "Frontier Design Benchmark", "Gestalt", "Claude Fable 5.2", "Claude Opus 5.5", "GPT-6.1 Sol", "Claude Sonnet 5.5", "Kimi K3.1 Pro"], "also_reported_by": [], "takeaway": "On Contra Labs' Frontier Design Benchmark, Gestalt outranks the human golden reference and Claude Fable 5.2 across 4,311 blind votes on 33 real client briefs, and the rubric's scores track designer ratings at r = 0.94.", "recommendation": {"advice": "Pair blind pairwise votes with a rubric scored on 35 dimensions; Contra Labs found rubric scores correlated with designer ratings at r = 0.94, making design quality explainable beyond raw Elo.", "playbook": "evals-testing", "playbook_url": "https://wpnews.pro/playbooks/evals-testing"}, "alternates": {"html": "https://wpnews.pro/news/two-stealth-models-just-appeared-on-the-leaderboard-both-ahead-of-fable-5-2", "markdown": "https://wpnews.pro/news/two-stealth-models-just-appeared-on-the-leaderboard-both-ahead-of-fable-5-2.md", "text": "https://wpnews.pro/news/two-stealth-models-just-appeared-on-the-leaderboard-both-ahead-of-fable-5-2.txt", "jsonld": "https://wpnews.pro/news/two-stealth-models-just-appeared-on-the-leaderboard-both-ahead-of-fable-5-2.jsonld"}}