multimodal benchmark
VLMsAreBlind
A vision-language benchmark that probes blind spots and brittle reasoning in multimodal models.
Updated Aug 7, 2026
Published models4
Registry coverage4
MetricScore
EvidenceB
VLMsAreBlind leaderboard
Sorted by the source-provided rank. Higher score is better according to the registry.
| 1 | AC | 97.0% | 100.0% | 4 | C | |
| 2 | AC | 97.0% | 66.7% | 4 | C | |
| 3 | AC | 96.9% | 33.3% | 4 | C | |
| 4 | AC | 96.7% | 0.0% | 4 | C |
Score distribution
Top published rows on the benchmark's original scale.
VLMsAreBlind highlights
The top published results on this benchmark's own scale.
What is VLMsAreBlind?
Definition and scoring fields from the benchmark registry.
A vision-language benchmark that probes blind spots and brittle reasoning in multimodal models.
Scores are shown in ratio. The current registry marks this benchmark as not independently verified with evidence level B.
- Family
- VLMsAreBlind
- Modality
- multimodal
- Primary category
- multimodal
- Score direction
- higher
- LLMBoard eligible
- No
- Evaluation key
- vlmsareblind|llm-stats-current
Source-native results are preserved. Eligibility for the overall LLMBoard score is a separate policy decision.
FAQ
Common questions about VLMsAreBlind.
Which model scores highest on VLMsAreBlind?
Qwen3.5-35B-A3B is currently ranked first with 97.0%.
What does VLMsAreBlind measure?
A vision-language benchmark that probes blind spots and brittle reasoning in multimodal models.
Is a higher score better?
Yes. Higher values rank better for this benchmark.
How many models are compared?
4 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.