Multimodal & Grounded benchmark
RealWorldQA leaderboard
Every model the catalog carries a published RealWorldQA value for, ranked by that value.
CategoryMultimodal & Grounded
MeasureImage-grounded QA
TasksReal-world visual question answering
DifficultyGeneral visual reasoning
Published byQwen3.6 launch benchmarks
A grounded visual QA benchmark focused on answering practical questions about real-world images and scenes.
RealWorldQA ranking
12 models with a published RealWorldQA value, ordered by that value, highest first. Models the source has not scored on this benchmark are not listed — they are unmeasured, not last.
| Rank | Model | Provider | Image-grounded QA |
|---|---|---|---|
| 1 |  Qwen3.8-Flash-Next | Alibaba | 88.5 |
| 2 | Qwen3.8 Max | Alibaba | 88.0 |
| 3 | Qwen3.8-Omni-Flash | Alibaba | 87.7 |
| 4 | Qwen3.7 Plus | Alibaba | 86.9 |
| 5 | Qwen3.8-27B | Alibaba | 85.9 |
| 6 | Qwen3.6-35B-A3B | Alibaba | 85.3 |
| 7 | Qwen3.6-27B | Alibaba | 84.1 |
| 8 | PMTernary Bonsai 2 27B | Prism ML | 80.1 |
| 9 |  LFM2.5-VL-3B | LiquidAI | 73.1 |
| 10 | ZYZAYA1-VL-8B | Zyphra | 65.0 |
| 11 |  North Micro Vision Instruct | Cohere | 62.2 |
| 12 | LFM2.5-VL-450M | LiquidAI | 58.4 |
Evidence key: Observed
Rows are ordered by the value Qwen3.6 launch benchmarks published, highest first. The source does not state whether a higher value is the better result, so this page does not either: for a benchmark that measures a rate of failure, read the table from the bottom.