MUSE: leaderboard
Large language models (LLMs) have recently advanced text-driven 3D generation, yet Text-to-CAD remains far from supporting industrial product design. Existing benchmarks focus primarily on generating single-part CAD models and evaluate them using geometric similarity metrics that fail to capture functionality, manufacturability, and assemblability. To address this gap, we introduce MUSE, a Text-to
Metric: Final Score (self-reported). Source: benchmarklist.com. Saturation forecast: Around May 2027. 15 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-5.5 | 67.14 | #31 |
| 2 | Gemini 3.1 Pro (Preview) | 53.93 | #58 |
| 3 | Claude Opus 4.7 | 53.46 | #49 |
| 4 | GLM-5.1 | 24.06 | #97 |
| 5 | Claude 3.7 Sonnet | 17.92 | #249 |
| 6 | Qwen 3.5 122B A10B | 10.38 | #173 |
| 7 | Qwen 2.5 72B | 10.38 | #499 |
| 8 | GPT-4o | 8.81 | #343 |
| 9 | MiniMax-M2.7 | 8.18 | #258 |
| 10 | Llama 3.1 70B | 5.5 | #592 |
| 11 | MiniMax-M2.5 | 5.35 | #303 |
| 12 | Qwen 3.6 35B A3B | 4.72 | #226 |
| 13 | GLM-4.7 Flash | 2.83 | #509 |
| 14 | Llama 3.1 8B | 0 | #1151 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=muse · How It Works · Data refreshed daily, snapshot 2026-10-20.