LongNovel (32K): leaderboard
Metric: Balanced accuracy (%; M, about 32k tokens context scale; binary detection of hallucinated summaries of Chinese novels and English BookSum chapters (8 hallucination types, human-revised test set); direct prompt with the summary at the end; an output without a valid consistency score counts as an incorrect prediction; chance is 50). Source: arxiv.org. Saturation forecast: Around December 2026. 12 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | DeepSeek V4 Flash | 81 |
| 2 | GPT-5.2 Instant | 77 |
| 3 | Claude Sonnet 4.5 | 66.5 |
| 4 | Qwen 3 8B | 60.8 |
| 5 | Qwen 3 14B | 59.8 |
| 6 | Qwen 3 32B | 57.5 |
| 7 | GLM-4 9B Chat | 51 |
| 8 | internlm2.5-20B-chat | 50.5 |
Interactive version: theaggregate.ai/benchmark?slug=longnovel-32k · How It Works · Data refreshed daily, snapshot 2026-09-29.