Massive Multimodal Embedding Benchmark for Video and Document Tasks

MMEB-V2 from TIGER-Lab is a benchmark dataset for evaluating multimodal AI models. It expands on its predecessor to include five new tasks focused on video and visual document analysis. The dataset was last updated in November 2025.

Use Cases

• Benchmark model performance on Video Retrieval and Video Classification tasks.
• Evaluate multimodal embeddings for Moment Retrieval and Video Question Answering.
• Test retrieval systems on the Visual Document Retrieval task.

Strengths

• Expands evaluation scope to include five new tasks, broadening benchmark coverage.
• Maintained by TIGER-Lab, a research organization, with a recent update in November 2025.

Limitations

• Specific dataset scale, such as the number of video clips or documents, is not provided.
• The raw description is truncated, limiting visibility into full task definitions and data composition.

Provenance

Source

TIGER-Lab

Collection Method

null

Time Range

null

Freshness

Last updated 2025-11-11.

Geography

null