MMLongBench-Write

MMLongBench-Write is a benchmark that contains six tasks designed to evaluate the long-generation capabilities of vision-language models using long output examples.

Papers using MMLongBench-Write (2)

IS-CoT: Breaking the Long-form Generation Collapse via Interleaved Structural Thinking2026LongWriter-V: Enabling Ultra-Long and High-Fidelity Generation in Vision-Language Models2025