DSV-Mem: leaderboard
Metric: Accuracy (%), 1,000 stateful visual-memory questions, mean of three runs. Source: arxiv.org. Saturation forecast: Around September 2027. 14 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Gemini 3.7 Flash (Medium) | 39.6 |
| 2 | Gemini 3.1 Pro (Preview) (Medium) | 27.5 |
| 3 | Kimi K3 (High) | 24.3 |
| 4 | Gemini 3.6 Flash (Medium) | 23 |
| 5 | MiniMax M3 (Thinking) | 13.1 |
| 6 | Qwen 3.5 397B A17B (Thinking) | 11.3 |
| 7 | Qwen 3.5 27B (Thinking) | 9.6 |
| 8 | Gemma 4 31B (IT) | 8.5 |
| 9 | Gemini 3.5 Flash Lite (Medium) | 8.2 |
| 10 | Qwen 3.5 9B (Thinking) | 7.2 |
| 11 | Kimi K2.5 (Thinking) | 5.8 |
| 12 | Gemma 4 26B A4B (IT) | 3.9 |
| 13 | Qwen 3.5 4B (Thinking) | 3.3 |
Interactive version: theaggregate.ai/benchmark?slug=dsv-mem · How It Works · Data refreshed daily, snapshot 2026-10-08.