AI Paper Atlas

Research map / Benchmarks / MSVD-QA

MSVD-QA leaderboard

81 papers at ICML, NeurIPS, ICLR, CVPR, AAAI, ICCV and ECCV, and the journals TPAMI, IJCV and JMLR report results on MSVD-QA. Best accuracy: 83.7 (LLaVA-Video + EchoBat, AAAI 2026).

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.

MSVD-QA · accuracy

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

#accuracy ↑MethodPaperTopic
183.7LLaVA-Video + EchoBat
LLaVA-Video-7B
EchoBat: Echo-Vision Enhancement and Echo-Layered Sampling for Video LLMs Hallucination Mitigation
AAAI 2026 · Table 2
Video-language understanding
280.36ISR-DPO (π θ9)
LLaVA-Next-7B-based VLMM
ISR-DPO: Aligning Large Multimodal Models for Videos by Iterative Self-Retrospective DPO
AAAI 2025 · Table 2
Multimodal LLMs
379.8VideoChat2* STEP
VideoChat2*
STEP: Enhancing Video-LLMs' Compositional Reasoning by Spatio-Temporal Graph-guided Self-Training
CVPR 2025 · Table 1
Video-language understanding
478.8TC-LLaVA
Vicuna-7B-v1.5
TC-LLaVA: Rethinking the Transfer of LLava from Image to Video Understanding with Temporal Considerations
AAAI 2025 · Table: Results of video questi
Video-language understanding
578.6DynImg
Qwen2.5-7B
DynImg: Key Frames with Visual Prompts are Good Representation for Multi-Modal Video Understanding
ICCV 2025 · Table 1
Video-language understanding
677.4DC+FreeVA
34B LLM
Dense Connector for MLLMs
NeurIPS 2024 · Table: Zero-shot Video QA
Multimodal LLMs
776.5MovieChat+MovieChat+: Question-Aware Sparse Memory for Long Video Question Answering
TPAMI 2025 · Table 1
Video-language understanding
875.9Slot-VLM†
Vicuna-7B
Slot-VLM: Object-Event Slots for Video-Language Modeling
NeurIPS 2024 · Table 1
Video-language understanding
975.8EMA
Qwen2-7B
Efficient Motion-Aware Video MLLM
CVPR 2025 · Table 1
Video-language understanding
1075.2AVCD
VideoLLaMA2
AVCD: Mitigating Hallucinations in Audio-Visual Large Language Models through Contrastive Decoding
NeurIPS 2025 · Table: Results on video-LLMs
Cross-modal learning
71 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.