# Design Composite Index

> Web apps, interface design, and visual layouts.

- Canonical URL: https://prochoice.industries/design
- Domain JSON endpoint: https://prochoice.industries/api/v1/domains/design
- Epoch version: `design-code-v1.0.0`
- Snapshot hash: `43237d0f0a7497fc`
- Updated: 2026-11-21
- Cost axis: Measured cost per Design task (Vals Vibe Code & DesignArena, USD, log scale)
- Evaluated entries: 92

## Headline picks

- Budget: Muse Ember at 58.1 and $0.029
- Best value: GPT-6.2 Sol at 72.9 and $2.18
- Top score: Claude Fable 5.2 at 81.1 and $24.7

## Pareto frontier

| Rank | Model | Effort | Provider | Score | Cost / task |
| :--- | :--- | :--- | :--- | :--- | :--- |
| 1 | Claude Fable 5.2 | Extra high | Anthropic | 81.1 | $24.7 |
| 2 | Claude Opus 5.5 | Max | Anthropic | 80.6 | $24.7 |
| 3 | Claude Sonnet 5.5 | Extra high | Anthropic | 75.1 | $19.8 |
| 4 | Gemini 4 Carbon | High | Google | 74.6 | $8.33 |
| 5 | Gemini 4 Neon | High | Google | 74.2 | $8.05 |
| 6 | GPT-6.2 Sol | Max | OpenAI | 72.9 | $2.18 |
| 7 | Muse Spark 1.3 | High | Meta | 69.8 | $1.13 |
| 8 | Muse Wonder | Max | Meta | 68.5 | $1.12 |
| 9 | Muse Spark 1.2 | High | Meta | 60.6 | $0.81 |
| 10 | Muse Ember | Max | Meta | 58.1 | $0.029 |

## Active benchmarks

- Vals Vibe Code v1.1 (1.1, ACTIVE): https://www.vals.ai/benchmarks/vibe-code
- LMArena WebDev (webdev, ACTIVE): https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset
- DesignArena Website Arena (public, ACTIVE): https://designarena.ai/leaderboard

Full JSON dataset: https://prochoice.industries/api/v1/snapshot
