VOLTBench - Mean Length Accuracy (100 Sections): leaderboard

Length adherence on VOLTBench (2026), a long-form generation benchmark of chapter-based story and code instructions, at the 100-section target: five generations are sampled per instruction and their mean length is scored as 100 x max(0, 1 - |mean - target| / target). Percent; higher is better.

Metric: Mean length accuracy (%): 100 x max(0, 1 - |mean output length - target length| / target length) over the five sampled generations per instruction, English simple-difficulty instructions, Story (unstructured) and Code Function (structured) chapter-based tasks, 100-section target; higher is better. Source: arxiv.org. Saturation forecast: Around October 2028. 9 models tracked. Updated 7 Oct 2026.

Top models

#ModelScoreOverall rank
1LongWriter-8B31.6
2DeepSeek V39.3#736
3Falcon3-Mamba-7B-Instruct6.5
4DeepSeek R16#484
5GPT-4o Mini4.8#1680
6Qwen 2.5 7B Instruct2.2#2732
7Llama 3.1 8B Instruct1.7#4655
8Claude 3.5 Sonnet0.9#755
9Qwen 2.5 1.5B Instruct0.7#7477

No result here: #2 Claude Fable 5.2, #3 Gemini 4 Neon, #4 GPT-6.2 Astra.

Interactive version: theaggregate.ai/benchmark?slug=voltbench-mean-length-accuracy-100-sections · How It Works · Data refreshed daily, snapshot 2026-11-22.