RealWorldQA: leaderboard
RealWorldQA: Evaluates multimodal understanding across image, text, chart, diagram, or cross-modal reasoning tasks.
Metric: RealWorldQA (self-reported). Source: benchmarklist.com. Status: years away from saturation. 48 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Qwen3.8-Flash-Next | 88.5 |
| 2 | Qwen 3.8 Flash | 88.5 |
| 3 | Qwen 3.8 Max | 88 |
| 4 | Qwen 3.7 Plus | 86.9 |
| 5 | Seed 2.1 Pro | 86.7 |
| 6 | Seed 2.1 Turbo | 86.3 |
| 7 | Qwen 3.8 27B | 85.9 |
| 8 | Qwen 3.6 Plus | 85.4 |
| 9 | Qwen 3.6 35B A3B | 85.3 |
| 10 | Qwen 3.5 122B A10B | 85.1 |
| 11 | Qwen 3.5 35B A3B | 84.1 |
| 12 | Qwen 3.6 27B | 84.1 |
| 13 | GPT-5.4 (xHigh) | 83.8 |
| 14 | Qwen 3.5 27B | 83.7 |
| 15 | Gemini 3.1 Pro (Preview) | 83.5 |
Interactive version: theaggregate.ai/benchmark?slug=realworldqa · How It Works · Data refreshed daily, snapshot 2026-09-05.