ZwZ2 Paper - MME-RealWorld (CN): leaderboard
MME-RealWorld Chinese split (Zhang et al.), a fine-grained perception benchmark in the paper's grouping: questions on high-resolution real-world images; not the original MME. ZwZ2 paper (arXiv 2610.37037v1, Table 1): the authors evaluated the base models Qwen3.5-4B, Qwen3.5-9B and Qwen3.6-27B and the ZwZ2-4B/9B/27B models they trained from them with GRPO in non-thinking mode. The table notes say only that all models were evaluated in low-thinking or non-thinking modes, without naming the mode per model; decoding, prompts, input resolution and evaluation framework are not reported. Accuracy (%), higher is better.
Metric: Accuracy (%). Source: arxiv.org. Saturation forecast: Around February 2027. 6 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Qwen 3.5 9B | 69.53 | #391 |
| 2 | Qwen 3.5 4B | 69.33 | #514 |
| 3 | Qwen 3.6 27B | 68.68 | #182 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=zwz2-paper-mme-realworld-cn · How It Works · Data refreshed daily, snapshot 2026-11-05.