ROSE: leaderboard
Multimodal large language models (MLLMs) are increasingly expected to act on visual information, yet the same scene may require different actions under different task contexts. How reliably can a model turn the same visual evidence into the action required by the current context? To answer this question, we introduce ROSE (Reference-conditioned Oddity and Sy
Metric: Avg. (self-reported). Source: benchmarklist.com. Saturation forecast: Around July 2027. 10 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-5.5 | 92.2 | #31 |
| 2 | Gemini 3.1 Pro (Preview) | 79.4 | #58 |
| 3 | Qwen 3.6 Plus | 50.3 | #120 |
| 4 | GLM-5V Turbo | 33.8 | #166 |
| 5 | Claude Opus 4.8 | 24.3 | #37 |
| 6 | Claude Sonnet 4.6 | 23.7 | #90 |
| 7 | GLM-4.6V | 20.7 | #316 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=rose · How It Works · Data refreshed daily, snapshot 2026-10-20.