Cookie-Bench: leaderboard
Front-end web code has become a core product surface for every frontier LLM release, yet evaluating these interactive applications at development speed remains costly because human-judged leaderboards like Arena do not scale. Existing automated proxies typically lean on reference implementations, test suites, or rigid checklists, and tend to miss the reasoned synthesis a human reviewer performs ov
Metric: React Overall (self-reported). Source: benchmarklist.com. Saturation forecast: Around August 2027. 13 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Claude Opus 4.7 | 83.3 | #49 |
| 2 | Claude Opus 4.6 | 80 | #63 |
| 3 | Kimi K2.6 | 78.3 | #104 |
| 4 | GPT-5.4 | 77.8 | #79 |
| 5 | MiMo-V2-Pro | 76.3 | #168 |
| 6 | GLM-5.1 | 76.2 | #97 |
| 7 | Qwen 3.6 Plus | 75.2 | #120 |
| 8 | Qwen 3.6 27B | 67.2 | #170 |
| 9 | Gemini 3.1 Pro (Preview) | 64.9 | #58 |
| 10 | Gemini 3 Flash | 63.3 | #99 |
| 11 | Qwen 3.6 35B A3B | 61.3 | #226 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=cookie-bench · How It Works · Data refreshed daily, snapshot 2026-10-20.