Benchmark Almanac

rednote hi lab (Xiaohongshu)

We track 2 benchmarks that rednote hi lab (Xiaohongshu) made or helped make. Its models are on its lab page.

0 in use 2 with public code 1 with a leaderboard here

Benchmarks

Name Kind Released Best score Status
VDR-Bench Models answer 2,000 visual questions that require finding entities in images and researching related facts. A Qwen3-VL judge checks final answers against references. Entity recall separately measures whether searches retrieve the relevant entities. Deep research Feb 2026 27.2% Not in cards
doteBench Models edit English and Chinese recordings using a transcript that marks where each change belongs. The evaluator checks edit success and preservation outside those regions. Audio understanding and voice Aug 2026 Not in cards