multimodal benchmark
ZEROBench
ZEROBench is a challenging vision benchmark designed to test models on zero-shot visual understanding tasks.
Updated Aug 7, 2026
Published models9
Registry coverage9
MetricScore
EvidenceB
ZEROBench leaderboard
Sorted by the source-provided rank. Higher score is better according to the registry.
| 1 | MA | 0.41 points | 100.0% | 9 | C | |
| 2 | ME | 0.33 points | 87.5% | 9 | C | |
| 3 | BY | 0.18 points | 75.0% | 9 | C | |
| 4 | MA | 0.11 points | 62.5% | 9 | C | |
| 5 | BY | 0.11 points | 50.0% | 9 | C | |
| 6 | AC | 0.1 points | 37.5% | 9 | C | |
| 7 | AC | 0.09 points | 25.0% | 9 | C | |
| 8 | AC | 0.08 points | 12.5% | 9 | C | |
| 9 | AC | 0.04 points | 0.0% | 9 | C |
Score distribution
Top published rows on the benchmark's original scale.
ZEROBench highlights
The top published results on this benchmark's own scale.
What is ZEROBench?
Definition and scoring fields from the benchmark registry.
ZEROBench is a challenging vision benchmark designed to test models on zero-shot visual understanding tasks.
Scores are shown in points. The current registry marks this benchmark as not independently verified with evidence level B.
- Family
- ZEROBench
- Modality
- image
- Primary category
- multimodal
- Score direction
- higher
- LLMBoard eligible
- No
- Evaluation key
- zerobench|llm-stats-current
Source-native results are preserved. Eligibility for the overall LLMBoard score is a separate policy decision.
FAQ
Common questions about ZEROBench.
Which model scores highest on ZEROBench?
Kimi K3 is currently ranked first with 0.41 points.
What does ZEROBench measure?
ZEROBench is a challenging vision benchmark designed to test models on zero-shot visual understanding tasks.
Is a higher score better?
Yes. Higher values rank better for this benchmark.
How many models are compared?
9 unique published model results are currently shown.
Does this benchmark affect the overall score?
This benchmark is preserved as source-native evidence but is not eligible for the current overall score.