Research map / Vision-language and multimodal / Cross-modal learning
NExT-OMNI: Towards Any-to-Any Omnimodal Foundation Models with Discrete Flow Matching
Run Luo, Xiaobo Xia, Lu Wang, Longze Chen, Renke Shan, Jing Luo, Min Yang, Tat-Seng Chua
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| replace AR paradigm with DFM (decoupled representation) (ablation) | NExT-OMNI baseline · Ablation average | average: 41.4 -> 42.6 | Improved |
| replace AR with DFM (ablation) | NExT-OMNI baseline · VQAv2 | accuracy: 55.2 -> 52.3 | Hurt |
| replace AR with DFM (ablation) | NExT-OMNI baseline · GenEval | GenEval: 53.4 -> 59.8 | Improved |
| replace AR with DFM (ablation) | NExT-OMNI baseline · Spoken QA (S->S) | accuracy: 16.4 -> 20.3 | Improved |
| use unified instead of decoupled representation (ablation) | DFM model · Ablation average | average: 42.6 -> 43.0 | Improved |
| use unified instead of decoupled representation (ablation) | DFM model · InfoSeek | Top-5 accuracy: 29.6 -> 32.8 | Improved |
| use unified instead of decoupled representation (ablation) | DFM model · VQAv2 | accuracy: 52.3 -> 51.7 | Hurt |
| add dynamic generation strategy (DGS) (ablation) | DFM unified model · VQAv2 | accuracy: 51.7 -> 54.3 | Improved |
| add dynamic generation strategy (DGS) (ablation) | DFM unified model · AudioCaps | score: 59.4 -> 61.9 | Improved |
| add dynamic generation strategy (DGS) (ablation) | DFM unified model · GenEval | GenEval: 59.2 -> 59.4 | No clear effect |
| add modality reconstruction loss (ablation) | DFM unified + DGS · GenEval | GenEval: 59.4 -> 62.6 | Improved |
| add modality reconstruction loss (ablation) | DFM unified + DGS · Ablation average | average: 43.9 -> 45.6 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 6 of 6 on most shared results.
- OpenOmni — OpenOmni: Advancing Open-Source Omnimodal Large Language Models with Progressive Multimodal Alignment and Real-time Emotional Speech Synthesis (NeurIPS 2025) · better on 2 of 3 shared results
- MMaDA — MMaDA: Multimodal Large Diffusion Language Models (NeurIPS 2025) · better on 2 of 2 shared results
- Janus — Janus: Decoupling Visual Encoding for Unified Multimodal Understanding and Generation (CVPR 2025) · better on 1 of 1 shared result
- VITA 1.5 — VITA-1.5: Towards GPT-4o Level Real-Time Vision and Speech Interaction (NeurIPS 2025) · better on 1 of 1 shared result
- FUDOKI — FUDOKI: Discrete Flow-based Unified Understanding and Generation via Kinetic-Optimal Velocities (NeurIPS 2025) · better on 1 of 1 shared result
- NExT-GPT — NExT-GPT: Any-to-Any Multimodal LLM (ICML 2024) · better on 1 of 1 shared result
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Omni-Diffusion: Unified Multimodal Understanding and Generation with Masked Discrete Diffusion (ICML 2026 · similarity 0.89)
- Omni-Interactive Universal Embedder (NeurIPS 2026 · similarity 0.88)
- OmniBench: Towards The Future of Universal Omni-Language Models (NeurIPS 2025 · similarity 0.88)
- OmniBind: Large-scale Omni Multimodal Representation via Binding Spaces (ICLR 2025 · similarity 0.88)
- OpenOmni: Advancing Open-Source Omnimodal Large Language Models with Progressive Multimodal Alignment and Real-time Emotional Speech Synthesis (NeurIPS 2025 · similarity 0.87)
- HEMM: Holistic Evaluation of Multimodal Foundation Models (NeurIPS 2024 · similarity 0.87)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
