LLM Stats (MMLongBench-Doc): leaderboard
Long-document multimodal benchmark testing understanding of lengthy PDFs and documents with interleaved text, figures, and tables.
Metric: Score (%). Source: llm-stats.com. Saturation forecast: Around January 2027. 7 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Nemotron 3.5 Super VL (BF16) | 62.82 | #165 |
| 2 | Qwen 3.6 Plus | 62 | #120 |
| 3 | Qwen 3.5 27B | 60.2 | #187 |
| 4 | Qwen 3.5 35B A3B | 59.5 | #256 |
| 5 | Qwen 3.5 122B A10B | 59 | #173 |
| 6 | Qwen 3 VL 235B A22B (Thinking) | 56.2 | #238 (Qwen 3 VL 235B A22B) |
| 7 | Seed 2.0 Mini | 48.9 | #276 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=llm-stats-mmlongbench-doc · How It Works · Data refreshed daily, snapshot 2026-10-20.