multimodal benchmark
VQAv2 (val) Leaderboard
VQAv2 is a balanced Visual Question Answering dataset containing open-ended questions about images that require understanding of vision, language, and commonsense knowledge to answer. VQAv2 addresses bias issues from the original VQA dataset by collecting complementary images such that every question is associated with similar images that result in different answers, forcing models to actually understand visual content rather than relying on language priors.
Updated Aug 17, 2026
Models3
Model coverage3
MetricScore
EvidenceB
VQAv2 (val) Ranking
Higher score ranks better on this benchmark.
Sort by
Rank | Model | Score | Percentile | Participants | Evidence | Evaluated |
|---|
| Rank01 | ModelGO | Score71.6% | Percentile100.0% | Participants3 | EvidenceC | Evaluated |
| Rank02 | ModelGO | Score71.0% | Percentile50.0% | Participants3 | EvidenceC | Evaluated |
| Rank03 | ModelGO | Score62.4% | Percentile0.0% | Participants3 | EvidenceC | Evaluated |
VQAv2 (val) Highlights
The leading models and scores on this benchmark.
VQAv2 (val) Score Distribution
A closer view of the leading scores on this benchmark.
The Top AI Models for VQAv2 (val)
The first five results on this benchmark, with official price and output speed added where the model identity can be matched.
Ranking basisThis vqav2 (val) AI model leaderboard uses descending score in the benchmark's original unit. The leaderboard ranking keeps matched price and speed data separate from benchmark evidence.
- 01GOGemma 3 12BGoogle
- Score
- 71.6%
- Speed
- Up to 33 tok/s via DeepInfra
Strengths
- Ranks #1 of 3 compared models
- 100th percentile on this benchmark
- C evidence result
Considerations
- This result measures VQAv2 (val), not total model capability
- 02GOGemma 3 27BGoogle
- Score
- 71.0%
- Speed
- Up to 33 tok/s via DeepInfra
Strengths
- Ranks #2 of 3 compared models
- 50th percentile on this benchmark
- C evidence result
Considerations
- This result measures VQAv2 (val), not total model capability
- 03GOGemma 3 4BGoogle
- Score
- 62.4%
- Speed
- Up to 33 tok/s via DeepInfra
Strengths
- Ranks #3 of 3 compared models
- 0th percentile on this benchmark
- C evidence result
Considerations
- This result measures VQAv2 (val), not total model capability
Selection summary
Best AI Models for VQAv2 (val)
Gemma 3 12B currently leads VQAv2 (val) with 71.6%. It is the top model on this specific benchmark, while the best LLM for the broader task should also be checked against other benchmarks, price and runtime.
Use this leaderboard with the supporting benchmark results and coverage details above. A leaderboard position summarizes the selected ranking signal; it does not replace workload-specific testing.
What is VQAv2 (val)?
What VQAv2 (val) measures and how its scores work.
VQAv2 is a balanced Visual Question Answering dataset containing open-ended questions about images that require understanding of vision, language, and commonsense knowledge to answer. VQAv2 addresses bias issues from the original VQA dataset by collecting complementary images such that every question is associated with similar images that result in different answers, forcing models to actually understand visual content rather than relying on language priors.
Scores are shown in ratio. This benchmark is not independently verified and has an evidence level of B.
- Family
- VQAv2 (val)
- Modality
- multimodal
- Primary category
- multimodal
- Score direction
- higher
- LLMBoard eligible
- Yes
- Evaluation key
- vqav2-(val)|llm-stats-current
Benchmark scores retain their original unit. Overall score eligibility is shown separately.
FAQ
Common questions about VQAv2 (val).
Which model scores highest on VQAv2 (val)?
Gemma 3 12B is currently ranked first with 71.6%.
What does VQAv2 (val) measure?
VQAv2 is a balanced Visual Question Answering dataset containing open-ended questions about images that require understanding of vision, language, and commonsense knowledge to answer. VQAv2 addresses bias issues from the original VQA dataset by collecting complementary images such that every question is associated with similar images that result in different answers, forcing models to actually understand visual content rather than relying on language priors.
Is a higher score better?
Yes. Higher values rank better for this benchmark.
How many models are compared?
3 model results are currently shown.
Does this benchmark affect the overall score?
Yes. This benchmark can contribute to the current LLMBoard capability score.