AI Paper Atlas

Research map / Benchmarks / WorldSense

WorldSense leaderboard

10 papers at ICML, NeurIPS, ICLR, CVPR, AAAI, ICCV and ECCV, and the journals TPAMI, IJCV and JMLR report results on WorldSense. Best accuracy: 65.8 (OmniVideo-R1, ICML 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.

WorldSense · accuracy

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

#accuracy ↑MethodPaperTopic
165.8OmniVideo-R1
Qwen3-Omni-30B-A3B
OmniVideo-R1: Reinforcing Audio-visual Reasoning with Query Intention and Modality Attention
ICML 2026 · Table 1
Cross-modal learning
253.73+ ASD-FT
Qwen2.5-Omni-7B
Who Says What: Symbolic Trimodal Binding Mechanisms in Audio-Visual LLMs
NeurIPS 2026 · Table 1
Cross-modal learning
348.9LatentOmni
Qwen2.5-Omni-7B
LatentOmni: Rethinking Omni-Modal Understanding via Unified Audio-Visual Latent Reasoning
NeurIPS 2026 · Table 1
Multimodal LLMs
447.2OmniAgent
Qwen2.5-Omni-7B
Native Active Perception as Reasoning for Omni-Modal Understanding
ICML 2026 · Table 2
Video-language understanding
546+ AVATAR
Qwen2.5-Omni-7B
AVATAR: Reinforcement Learning to See, Hear, and Reason Over Video
CVPR 2026 · Table 1
Video-language understanding
645.84AV-Reasoner
Ola-Omni-7B
AV-Reasoner: Improving and Benchmarking Clue-Grounded Audio-Visual Counting for MLLMs
CVPR 2026 · Table 3
Video-language understanding
745EchoingPixels (20% tokens)
Qwen2.5-Omni-3B
EchoingPixels: Aliasing-Resistant Joint Token Reduction for Audio-Visual LLMs
ICML 2026 · Table 1
Cross-modal learning
844.1Ours (base Qwen2.5)
Qwen2.5-14B controller
Training-Free Multimodal Large Language Model Orchestration
ICML 2026 · Table 1
Cross-modal learning
940.22+OFF (FutureOmni-Train)
Qwen2.5-Omni-7B
FutureOmni: Evaluating Future Forecasting from Omni-Modal Context for Multimodal LLMs
ICML 2026 · Table 6
Multimodal LLMs
1038.5Confidence-aware voting (N=8)
Qwen2.5-VL-7B
Open-o3-Video: Grounded Video Reasoning with Explicit Spatio-Temporal Evidence
ICML 2026 · Table 11
Video-language understanding

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