Loading page…
Preferred languages
More languages supported by Google Translate
Open the selector to load the current language catalog.
Translations are provided by Google Translate.
Loading page…
Preferred languages
More languages supported by Google Translate
Open the selector to load the current language catalog.
Translations are provided by Google Translate.
external benchmark
BlenderWorld: Benchmarking Visual Agents for 3D Creation from Real-World References (text-only subset). Every model the catalog carries a published BlenderWorld (text-only) value for, ranked by that value.
The 78 text-brief tasks of BlenderWorld, a benchmark where agents work inside Blender to turn real-world 3D creation requests into editable Blender projects and requested exports. 7 models in the ModelScale catalog have a published value.
Each point is a model with a published BlenderWorld (text-only) value and complete token prices, on the 75% input / 25% output blend. Frontier models are spaced evenly along the cost axis.
7 models with a published BlenderWorld (text-only) value, ordered by that value, highest first. Models the source has not scored on this benchmark are not listed — they are unmeasured, not last.
| Rank | Model | Provider | Task score (0-100), half rubric and half visual quality |
|---|---|---|---|
| 1 | Anthropic | 86.1 | |
| 2 | OpenAI | 85.3 | |
| 3 | DeepSeek |
What the benchmark measures and how its value is published, from the source’s own record.
BlenderWorld (text-only) (BlenderWorld: Benchmarking Visual Agents for 3D Creation from Real-World References (text-only subset)) is an external benchmark published by BlenderWorld: Benchmarking Visual Agents for 3D Creation from Real-World References. The 78 text-brief tasks of BlenderWorld, a benchmark where agents work inside Blender to turn real-world 3D creation requests into editable Blender projects and requested exports. Its evaluation set is described by the source as 78 text-brief 3D creation tasks in Blender. The source rates its difficulty as long-horizon 3D modeling and scene building.
Each model's result is published by BlenderWorld: Benchmarking Visual Agents for 3D Creation from Real-World References as task score (0-100), half rubric and half visual quality. ModelScale prints that value to 1 decimal place, exactly as published, and never converts or rescales it. This page ranks the 7 models that carry a published value, highest value first; a model the source never scored is left out rather than ranked at zero. The source does not state whether a higher value is the better result, so neither does this page.
| 4 | Baidu | 79.0 |
| 5 | Tencent | 76.8 |
| 6 | Z.AI | 71.4 |
| 7 | Z.AI | 69.7 |
Evidence key: Observed
Rows are ordered by the value BlenderWorld: Benchmarking Visual Agents for 3D Creation from Real-World References published, highest first. The source does not state whether a higher value is the better result, so this page does not either: for a benchmark that measures a rate of failure, read the table from the bottom.