Multimodal · Benchmark profile
Visual reasoning stress test designed so current frontier models score near zero, measuring progress on hard multimodal puzzles.
Data verified 27 Jul 2026 · Methodology 1.8.0
Benchmark score on ZeroBench
GPT-5.4 leads at 41%, followed by Muse Spark (33%) and Gemini 3.1 Pro (29%).
4 ranked models · higher is better · labels show rank and score
| Model | Rank | Provider | Score |
|---|---|---|---|
| GPT-5.4 | #1 | OpenAI | 41% |
| Muse Spark | #2 | Meta | 33% |
| Gemini 3.1 Pro | #3 | 29% | |
| Kimi K3 | #4 | Moonshot AI | 23% |
One best score per model · higher is better
| Rank | Model | Provider | License | Evidence use | Score |
|---|---|---|---|---|---|
| #1 | GPT-5.4 gpt-5.4 | OpenAI | closed | Reference only | 41% |
| #2 | Muse Spark muse-spark | Meta | closed | Reference only | 33% |
| #3 | Gemini 3.1 Pro gemini-3-1-pro | closed | Estimated reference | 29% | |
| #4 | Kimi K3 kimi-k3 | Moonshot AI | closed | Reference only | 23% |
The top of this snapshot is led by GPT-5.4 at 41%; third place is 12.0 points behind. The top-4 spread is 18.0 points.
About ZeroBench
Visual reasoning stress test designed so current frontier models score near zero, measuring progress on hard multimodal puzzles. Results stay tied to the exact model variant and evaluation system. Multiple systems for the same model use the best published score on this page; overall Lumina scoring uses the median of ranking-eligible rows.
Open benchmark source ↗FAQ
Visual reasoning stress test designed so current frontier models score near zero, measuring progress on hard multimodal puzzles.
GPT-5.4 by OpenAI currently leads with 41%.
4 models in the LuminaBench cohort have a qualifying score on this benchmark.
No. This benchmark is display-only and does not enter the overall Lumina composite.
Related
