AI Paper Atlas

Research map / Benchmarks / MLVU M-Avg

MLVU M-Avg leaderboard

17 papers at ICML, NeurIPS, ICLR, CVPR, AAAI, ICCV and ECCV, and the journals TPAMI, IJCV and JMLR report results on MLVU M-Avg. Best accuracy: 83.9 (LVAgent, ICCV 2025).

Numbers are as each paper reports them, read by AI from its results tables and not checked by hand. Papers can differ in setting (training data, model size, extra data); the topic column shows what each paper is about. Open the paper before you compare or quote a number.

MLVU M-Avg · accuracy

17 papers, each with its best reported number for its own method; higher is better.

#accuracy ↑MethodPaperTopic
183.9LVAgentLVAgent: Long Video Understanding by Multi-Round Dynamical Collaboration of MLLM Agents
ICCV 2025 · Table 1
Video-language understanding
283.6META
Qwen3-VL-8B
META: Meta Evolution of Tool Trajectory Adaptation for Long-Video Understanding
CVPR 2026 · Table 1
Video-language understanding
376.2ReQuest
Qwen3-VL-8B
ReQuest: Rethinking-based Question-Aware Frame Selection for Long-Form Video QA
ECCV 2026 · Table 1
Video-language understanding
475.1MiniCPM-V 4.5MiniCPM-V 4.5: Cooking Efficient MLLMs via Architecture, Data, and Training Recipe
CVPR 2026 · Table 1
Multimodal LLMs
574.7InternVideo2.5* + LongVPO Stage2-iter2
InternVideo2.5-8B
LongVPO: From Anchored Cues to Self-Reasoning for Long-Form Video Preference Optimization
NeurIPS 2025 · Main results table
Video-language understanding
673.38Ours
Llava-Video-7B
From Trial to Triumph: Advancing Long Video Understanding via Visual Context Sample Scaling and Self-reward Alignment
ICCV 2025 · Table 1
Video-language understanding
771.6NVILA-8B-Video + AutoGaze (1024 frames, 3584 res)
NVILA-8B-Video
Attend Before Attention: Efficient and Scalable Video Understanding via Autoregressive Gazing
CVPR 2026 · Table 1
Video-language understanding
870.9DynamicSelect
Qwen2.5-VL-7B
Seeing Is Believing: Grounding Long-Video Understanding in Spatio-Temporal Visual Evidence
AAAI 2026 · Table 3
Video-language understanding
970.5LiVi-LLM-7BLiViBench: An Omnimodal Benchmark for Interactive Livestream Video Understanding
AAAI 2026 · Table 3
Video-language understanding
1069.8Reflect-R1 (Ours)Reflect-R1: Evidence-Driven Reflection for Self-Correction in Long Video Understanding
ECCV 2026 · Table 1
Video-language understanding
7 more papers on this leaderboard, the baselines reported on it and a CSV download are free with a scime account.
Sign in free to see all

Not affiliated with any conference, journal or the benchmark's authors. Corrections: contact@latenticlabs.com.