A Benchmark for Visual Question Answering using World Knowledge (A-OKVQA)
We show this table for reference; we do not rank on it.
A visual question answering benchmark whose questions cannot be answered from the image alone and require commonsense or world knowledge.
Benchmark score on A-OKVQA — November 13, 2026
We compile the A-OKVQA rows from provider self-reports. Ternary Bonsai 2 27B leads the table at 86.8%. We do not use these results to rank models overall.
1 modelMultimodal & GroundedStaleDisplay onlyUpdated November 13, 2026
Benchmark score table (1 model)
The one result on this page was reported by the lab that built the model.
| Rank | Model / configuration | Score | Who ran it | Parameters (B) | Open / closed |
|---|---|---|---|---|---|
| 1 | Ternary Bonsai 2 27BPrism MLReported by Prism ML | 86.8% | Reported by Prism ML | Not reported | Open |
- Lab-reported
- Benchmark-run
- Independent
About A-OKVQA
Year
2022
Tasks
Knowledge-grounded visual question answering
Format
Multiple choice and direct answer
Difficulty
Commonsense and world knowledge about images
BenchLM stores A-OKVQA as a display-only knowledge-grounded visual QA reference outside the weighted core schema. Providers report either the multiple-choice or the direct-answer setting, so the row note should record which split and setting produced the value.
Freshness and provenance
Version
A-OKVQA 2022
Refresh cadence
Static
Staleness state
Stale
Question availability
Public benchmark set
BenchLM uses freshness metadata to decide whether a benchmark should still be treated as a strong differentiator, a benchmark to watch, or a display-only reference. For the full scoring policy, see the BenchLM methodology page.
Questions
What does A-OKVQA measure?
A visual question answering benchmark whose questions cannot be answered from the image alone and require commonsense or world knowledge.
Which model scores highest on A-OKVQA?
Ternary Bonsai 2 27B by Prism ML currently leads with a score of 86.8% on A-OKVQA.
How many models are evaluated on A-OKVQA?
1 AI models have published results on A-OKVQA in the BenchLM catalog.
Know when it’s worth switching models
The model to choose, the cheaper alternative, and the release we would wait on.
Read a sample issueJoin 7,000+ subscribers.
One email each week. Unsubscribe anytime.