AI Paper Atlas

Research map / Benchmarks / NExT-GQA

NExT-GQA leaderboard

16 papers at ICML, NeurIPS, ICLR, CVPR, AAAI, ICCV and ECCV, and the journals TPAMI, IJCV and JMLR report results on NExT-GQA. Best mIoU: 43 (VITAL-7B, CVPR 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.

NExT-GQA · mIoU

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

#mIoU ↑MethodPaperTopic
143VITAL-7B
Qwen2.5-VL-7B
Thinking With Videos: Multimodal Tool-Augmented Reinforcement Learning for Long Video Reasoning
CVPR 2026 · Table 4
Video-language understanding
236.8DeepVideo-R1-7BDeepVideo-R1: Video Reinforcement Fine-Tuning via Difficulty-aware Regressive GRPO
NeurIPS 2025 · Table 3
Instruction and fine-tuning
333.4VideoTemp-o3-7B-RL
Qwen2.5-VL-7B
VideoTemp-o3: Harmonizing Temporal Grounding and Video Understanding in Agentic Thinking-with-Videos
ICML 2026 · Table 3
Video-language understanding
432.9APPO
Qwen2.5-VL-7B
APPO: Attention-guided Perception Policy Optimization for Video Reasoning
CVPR 2026 · Table 1
Video-language understanding
528.1Qwen3-VL-4B + SynRL
Qwen3-VL-4B
Learning Transferable Temporal Primitives for Video Reasoning via Synthetic Videos
CVPR 2026 · Table 3
Video-language understanding
624.4TOGA
Mistral-7B
TOGA: Temporally Grounded Open-Ended Video QA with Weak Supervision
ICCV 2025 · Table 1
Video-language understanding
723.4MSR-ViRL
LLaVA-NeXT
Modularized Self-Reflected Video Reasoner for Multimodal LLM with Application to Video Question Answering
ICML 2025 · Table 4
Video-language understanding
819.3VideoStreaming
Vicuna-7B + Phi-2 encoder
Streaming Long Video Understanding with Large Language Models
NeurIPS 2024 · Next-GQA results table
Video-language understanding
917.5GroundFormerWhat You Ask is What You Ground: Bridging Question Intent to Temporal Evidence for Grounded VideoQA
ECCV 2026 · Table 2
Visual question answering
1015.6TimeCraft
Temp[CLIP]
TimeCraft: Navigate Weakly-Supervised Temporal Grounded Video Question Answering via Bi-directional Reasoning
ECCV 2024 · Table 1
Video-language understanding
1 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

NExT-GQA · Acc@GQA

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

#Acc@GQA ↑MethodPaperTopic
128.2VideoMind
Qwen2-VL-7B
VideoMind: A Chain-of-LoRA Agent for Temporal-Grounded Video Reasoning
ICLR 2026 · Table 4
Video-language understanding
228.1DisTime-InternVL
InternVL2.5-8B
DisTime: Distribution-based Time Representation for Video Large Language Models
ICCV 2025 · Table 6
Video-language understanding
327.6VITEDVITED: Video Temporal Evidence Distillation
CVPR 2025 · Table 2
Video-language understanding
424.6TOGA
Mistral-7B
TOGA: Temporally Grounded Open-Ended Video QA with Weak Supervision
ICCV 2025 · Table 1
Video-language understanding
521.5GroundFormerWhat You Ask is What You Ground: Bridging Question Intent to Temporal Evidence for Grounded VideoQA
ECCV 2026 · Table 2
Visual question answering
619WICIWavelet-based Intra-video Counterfactual Reasoning for Video Question Grounding
ECCV 2026 · Table 1
Video-language understanding
718.8CRA (FrozenBiLM)
FrozenBiLM
Cross-modal Causal Relation Alignment for Video Question Grounding
CVPR 2025 · Table 1
Visual question answering
818.6MSR-ViRL
LLaVA-NeXT
Modularized Self-Reflected Video Reasoner for Multimodal LLM with Application to Video Question Answering
ICML 2025 · Table 4
Video-language understanding
918.5TimeCraft
FrozenBiLM
TimeCraft: Navigate Weakly-Supervised Temporal Grounded Video Question Answering via Bi-directional Reasoning
ECCV 2024 · Table 1
Video-language understanding
1017.8VideoStreaming
Vicuna-7B + Phi-2 encoder
Streaming Long Video Understanding with Large Language Models
NeurIPS 2024 · Next-GQA results table
Video-language understanding
1 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.