SimpleVQA
We show this table for reference; we do not rank on it.
A visual question answering benchmark focused on straightforward image-grounded understanding.
Benchmark score on SimpleVQA — November 13, 2026
We compile the SimpleVQA rows from provider self-reports and secondary reports. Qwen3.7 Plus leads the table at 81.7%, followed by Step 3.7 Flash (79.2%) and Qwen3.8 Max (75.0%). We do not use these results to rank models overall.
Qwen3.7 Plus
Alibaba
Reported by AlibabaStep 3.7 Flash
StepFun
Reported by StepFunQwen3.8 Max
Alibaba
Reported by Alibaba18 modelsMultimodal & GroundedCurrentDisplay onlyUpdated November 13, 2026
Benchmark score table (18 models)
None of the 18 results on this page has a documented run by the benchmark or an independent evaluator: 12 reported by the labs that built the models and 6 reported in another publisher’s document.
| Rank | Model / configuration | Score | Who ran it | Parameters (B) | Open / closed |
|---|---|---|---|---|---|
| 1 | Qwen3.7 PlusAlibabaReported by Alibaba | 81.7% | Reported by Alibaba | Not reported | Closed |
| 2 | Step 3.7 FlashStepFunReported by StepFun | 79.2% | Reported by StepFun | Not reported | Open |
| 3 | Qwen3.8 MaxAlibabaReported by Alibaba | 75.0% | Reported by Alibaba | Not reported | Open |
| 4 | Muse Spark OpenMetaReported by Meta | 73.6% | Reported by Meta | Not reported | Open |
| 5 | Gemini 3 ProGoogleAnother lab’s comparison table | 73.2% | Another lab’s comparison table | Not reported | Closed |
| 6 | dots3-note PreviewDots StudioReported by Dots Studio | 72.5% | Reported by Dots Studio | Not reported | Open |
| 7 | Gemini 3.1 ProGoogleReported by Meta AI (comparison table) | 72.4% | Reported by Meta AI (comparison table) | Not reported | Closed |
| 8 | Muse SparkMetaReported by Meta | 71.3% | Reported by Meta | Not reported | Closed |
| 9 | Kimi K2.5Moonshot AIReported by Moonshot AI | 71.2% | Reported by Moonshot AI | Not reported | Open |
| 10 | Qwen3.5 397BAlibabaReported by Alibaba | 67.1% | Reported by Alibaba | Not reported | Open |
| 11 | Claude Opus 4.5AnthropicAnother lab’s comparison table | 65.7% | Another lab’s comparison table | Not reported | Closed |
| 12 | GPT-5.4OpenAIReported by Meta AI (comparison table) | 61.1% | Reported by Meta AI (comparison table) | Not reported | Closed |
| 13 | Qwen3.6-35B-A3BAlibabaReported by Alibaba | 58.9% | Reported by Alibaba | Not reported | Open |
| 14 | Grok 4.20xAIReported by Meta AI (comparison table) | 57.4% | Reported by Meta AI (comparison table) | Not reported | Closed |
| 15 | Qwen3.6-27BAlibabaReported by Alibaba | 56.1% | Reported by Alibaba | Not reported | Open |
| 16 | GPT-5.2OpenAIAnother lab’s comparison table | 55.8% | Another lab’s comparison table | Not reported | Closed |
| 17 | LFM3-VL-2BLiquidAIReported by LiquidAI | 36.3% | Reported by LiquidAI | Not reported | Open |
| 18 | LFM2.5-VL-3BLiquidAIReported by LiquidAI | 35.4% | Reported by LiquidAI | Not reported | Open |
- Lab-reported
- Benchmark-run
- Independent
Among the reported SimpleVQA rows, Qwen3.7 Plus is first at 81.7%. The third row is 6.7 points behind. The broader top-10 range is 14.6 points, so the table still separates the published systems.
18 models have been evaluated on SimpleVQA. The benchmark falls in the Multimodal & Grounded category. SimpleVQA is currently displayed for reference but excluded from the scoring formula, so it does not directly affect overall rankings.
About SimpleVQA
Year
2026
Tasks
Visual QA tasks
Format
Image-grounded question answering
Difficulty
General visual understanding
BenchLM uses SimpleVQA as a display-only visual QA reference rather than a weighted multimodal ranking input.
Freshness and provenance
Version
SimpleVQA 2026
Refresh cadence
Quarterly
Staleness state
Current
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 SimpleVQA measure?
A visual question answering benchmark focused on straightforward image-grounded understanding.
Which model scores highest on SimpleVQA?
Qwen3.7 Plus by Alibaba currently leads with a score of 81.7% on SimpleVQA.
How many models are evaluated on SimpleVQA?
18 AI models have published results on SimpleVQA in the BenchLM catalog.
Compare top models on SimpleVQA
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.