VSysBench - Task Accuracy: leaderboard
Metric: Task accuracy (%; mean judged answer score against the MM-Vet v2 ground truth while the system constraint applies, both halves; VSysBench: 2,258 verified MM-Vet v2 image questions, each under a system-message constraint from 22 sub-categories in 5 categories, once with an aligned and once with a conflicting user message (4,516 instances); default inference settings; GPT-5-mini judge at temperature 0). Source: arxiv.org. Saturation forecast: Around 2029. 16 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Claude Opus 4.7 | 45.1 |
| 2 | Claude Sonnet 4.6 | 44.2 |
| 3 | Qwen 3 VL 32B | 43.8 |
| 4 | GPT-5.4 | 42.7 |
| 5 | GPT-5.4 Mini | 37.9 |
| 6 | GPT-4o | 36.1 |
| 7 | Claude Haiku 4.5 | 34.7 |
| 8 | InternVL3.5-8B | 34.7 |
| 9 | GPT-5.4 Nano | 28.8 |
| 10 | Phi-4 Multimodal Instruct | 20.3 |
Interactive version: theaggregate.ai/benchmark?slug=vsysbench-task-accuracy · How It Works · Data refreshed daily, snapshot 2026-09-29.