Benchmarks
351 sources · 1,261 models × 3,053 benchmarks · 160,189 scoresVibePrevisBench's deterministic half: rule-based checks of each exported Blender scene against the task's hidden ground-truth facts, averaged over the 242 tasks, from the same runs as the overall score.
163rd strongest of 803 fitted benchmarks
Signal-to-noise +11.05 for the VibePrevisBench family
Another benchmark represents this family in the fit.
| # | Model | Setting | Score | Source |
|---|---|---|---|---|
| 1 | GPT-6 Astra | max · Codex | 70.0 | |
| 2 | GPT-6.1 Sol | max · Codex | 69.0 | |
| 3 | Claude Opus 5.5 | high · Claude Code | 68.4 | |
| 4 | Gemini 4 Argon | high · Claude Code | 68.4 | |
| 5 | GPT-5.6 Sol | max · Codex | 62.9 | |
| 6 | DeepSeek V4.1 Flash 1017 | max · Claude Code | 61.3 | |
| 7 | Gemini 3.8 Flash | high · Claude Code | 60.7 | |
| 8 | GLM-5.3 Flash | high · Claude Code | 57.8 | |
| 9 | Muse Spark 1.3 | x-high · Claude Code | 57.2 |