Research map / Vision-language and multimodal / Multimodal LLMs
Query-Kontext: An Unified Multimodal Model for Image Generation and Editing
YuXin Song, Wenkai Dong, Shizun Wang, Qi Zhang, Song Xue, Tao Yuan, Hu Yang, Haocheng Feng, Hang Zhou, Xinyan Xiao, Jingdong Wang
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 |
|---|---|---|---|
| use Query-Kontext instead of Qwen-Image on instruction editing | Query-Kontext · GEdit-Bench-EN | G_O: 7.56 -> 7.66 | Improved |
| use Query-Kontext instead of Qwen-Image on instruction editing | Query-Kontext · GEdit-Bench-CN | G_O: 7.52 -> 7.65 | Improved |
| use Query-Kontext instead of Qwen-Image for perceptual quality | Query-Kontext · GEdit-Bench-EN | G_PQ: 7.86 -> 7.37 | Hurt |
| use Query-Kontext instead of UNO-FLUX on DreamBooth | Query-Kontext · DreamBooth | DINO: 0.760 -> 0.786 | Improved |
| use Query-Kontext instead of Metaquery on DreamBooth | Query-Kontext · DreamBooth | CLIP-I: 0.851 -> 0.858 | Improved |
| use Query-Kontext instead of OmniGen on DreamBooth | Query-Kontext · DreamBooth | CLIP-T: 0.315 -> 0.307 | Hurt |
| use Query-Kontext instead of OmniGen on DreamBench | Query-Kontext · DreamBench | CLIP-T: 0.331 -> 0.336 | Improved |
| use Query-Kontext instead of UNO-FLUX on DreamBench | Query-Kontext · DreamBench | DINO: 0.542 -> 0.532 | Hurt |
| use Query-Kontext instead of BAGEL on GenEval | Query-Kontext · GenEval | Overall: 0.88 -> 0.88 | No clear effect |
| condition diffusion model on text plus Query-Kontext tokens from fine-tuned VLM instead of LLM text embeddings only (ablation) | Query-Kontext Stage 2 · Diffusion alignment | convergence speed | Improved |
| increase LoRA rank across 64, 128, 256 (ablation) | Query-Kontext · LoRA rank study | convergence: faster convergence at higher ranks, marginal quality gains beyond r=128 | Mixed |
| unfreeze the diffusion model in Stage 2 (ablation) | Query-Kontext · Stage 2 alignment (0.9B/4B/10B) | alignment success | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- OmniGen2 — OmniGen2: Towards Instruction-Aligned Multimodal Generation (CVPR 2026) · 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
- DiffThinker: Towards Generative Multimodal Reasoning with Diffusion Models (ICML 2026 · similarity 0.92)
- WEAVE: Unleashing and Benchmarking the In-context Interleaved Comprehension and Generation (CVPR 2026 · similarity 0.90)
- Generative Multimodal Pretraining with Discrete Diffusion Timestep Tokens (CVPR 2025 · similarity 0.90)
- Janus-Pro-R1: Advancing Collaborative Visual Comprehension and Generation via Reinforcement Learning (NeurIPS 2025 · similarity 0.90)
- ROVER: Benchmarking Reciprocal Cross-Modal Reasoning for Omnimodal Generation (ICLR 2026 · similarity 0.90)
- DuoGen: Towards Autonomous Interleaved Multimodal Generation (CVPR 2026 · similarity 0.90)
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.
