Benchmarks · creative
ClutterBench
Clutter as a share of drawn area in generated graphics, using the raw prompt. Red area plus half the orange area, averaged over 6 scored posts. Lower is better.
Category
11 models · 40 results · updated 2026-10-19 · clutterbench.adamholter.chatgpt.site ↗
Leaderboard
Sonnet 5.5 low
All results
40 rows| # | Model | Effort | ClutterBench |
|---|---|---|---|
| 1 | Sonnet 5.5 | low | 28.8% |
| 2 | Opus 5.5 | low | 31.3% |
| 3 | GPT-6 Astra | xhigh | 33.2% |
| 4 | Grok 4.7 | high | 33.8% |
| 5 | Sonnet 5.5 | medium | 35.0% |
| 6 | Sonnet 5.5 | high | 36.5% |
| 7 | Haiku 5.5 | xhigh | 37.8% |
| 8 | Sonnet 5.5 | xhigh | 40.1% |
| 9 | DeepSeek V4.1 Flash | xhigh | 40.8% |
| 10 | Opus 5.5 | xhigh | 41.7% |
| 11 | Haiku 5.5 | low | 41.8% |
| 12 | Grok 4.7 | low | 42.5% |
| 13 | GPT-6.1 Sol | medium | 42.7% |
| 14 | GPT-6 Astra | high | 43.0% |
| 15 | DeepSeek V4.1 Flash | low | 43.3% |
| 16 | GPT-6.1 Sol | low | 45.6% |
| 17 | GPT-6 Astra | low | 45.8% |
| 18 | GPT-6.1 Sol | high | 45.8% |
| 19 | Haiku 5.5 | medium | 46.3% |
| 20 | DeepSeek V4.1 Flash | high | 46.3% |
| 21 | Grok 4.7 | xhigh | 46.4% |
| 22 | DeepSeek V4.1 Flash | max | 46.6% |
| 23 | DeepSeek V4.1 Flash | medium | 46.8% |
| 24 | GPT-6 Astra | medium | 47.7% |
| 25 | Opus 5.5 | medium | 47.9% |
| 26 | Haiku 5.5 | high | 48.9% |
| 27 | Opus 5.5 | high | 48.9% |
| 28 | GPT-6.1 Sol | xhigh | 52.1% |
| 29 | GPT-6 Luna | high | 53.5% |
| 30 | Grok 4.7 | medium | 53.9% |