multimodal benchmark
BrowseComp-VL
BrowseComp-VL is the vision-language variant of BrowseComp, evaluating multimodal models on web browsing comprehension tasks that require processing visual web page content alongside text.
Updated Aug 11, 2026
Models1
Model coverage1
MetricScore
EvidenceB
BrowseComp-VL Ranking
Higher score ranks better on this benchmark.
| 01 | ZA | 51.9% | 100.0% | 1 | C |
BrowseComp-VL Highlights
The leading models and scores on this benchmark.
What is BrowseComp-VL?
What BrowseComp-VL measures and how its scores work.
BrowseComp-VL is the vision-language variant of BrowseComp, evaluating multimodal models on web browsing comprehension tasks that require processing visual web page content alongside text.
Scores are shown in ratio. This benchmark is not independently verified and has an evidence level of B.
- Family
- BrowseComp-VL
- Modality
- multimodal
- Primary category
- multimodal
- Score direction
- higher
- LLMBoard eligible
- No
- Evaluation key
- browsecomp-vl|llm-stats-current
Benchmark scores retain their original unit. Overall score eligibility is shown separately.
FAQ
Common questions about BrowseComp-VL.
Which model scores highest on BrowseComp-VL?
GLM-5V-Turbo is currently ranked first with 51.9%.
What does BrowseComp-VL measure?
BrowseComp-VL is the vision-language variant of BrowseComp, evaluating multimodal models on web browsing comprehension tasks that require processing visual web page content alongside text.
Is a higher score better?
Yes. Higher values rank better for this benchmark.
How many models are compared?
1 model results are currently shown.
Does this benchmark affect the overall score?
No. This benchmark is shown for reference but does not contribute to the overall score.