Mage-Flow: An Efficient Native-Resolution Foundation Model for Image Generation and Editing

• URL: http://arxiv.org/abs/2607.19064v2
• Date: Wed, 22 Jul 2026 15:23:32 GMT
• Title: Mage-Flow: An Efficient Native-Resolution Foundation Model for Image Generation and Editing
• Abstract summary: We introduce Mage-Flow, a compact 4B-scale generative stack for efficient text-to-image generation and instruction-based image editing.<n> stack is built from two co-designed components: Mage-VAE, a lightweight high-fidelity latent tokenizer, and a Native-Resolution Multi-Resolution Diffusion Transformer.<n>We develop a complete model family with Base, RL-aligned, and Turbo variants for both generation and editing.
• Score: 57.690762576837045
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Large-scale visual generators are increasingly capable but costly to train, fine-tune, and deploy. We introduce Mage-Flow, a compact 4B-scale generative stack for efficient text-to-image generation and instruction-based image editing. The stack is built from two co-designed components: Mage-VAE, a lightweight high-fidelity latent tokenizer, and a Native-Resolution Multimodal Diffusion Transformer trained with rectified flow matching. Mage-VAE uses one-step diffusion-style encoding and decoding with anchor-latent regularization, preserving the reconstruction quality of strong public VAEs while reducing tokenization cost by more than an order of magnitude. Together with native-resolution packing and stack-level CUDA kernel fusion, the stack supports flexible-resolution training and improves end-to-end training throughput by about $2.5\times$. Built on this foundation, we develop a complete model family with Base, RL-aligned, and Turbo variants for both generation and editing. Diffusion-NFT improves prompt following, text rendering, aesthetic quality, and editing fidelity, while few-step distillation with adversarial perceptual guidance produces 4-step Turbo models for low-latency inference. Despite its compact scale, Mage-Flow and Mage-Flow-Edit achieves competitive performance across standard generation and editing benchmarks. More importantly, the Turbo variants make high-resolution generation and editing practical for interactive use: at $1024^2$ resolution on a single NVIDIA A100 GPU, Mage-Flow-Turbo generates an image in 0.59s, and Mage-Flow-Edit-Turbo edits an image in 1.02s, while maintaining a small memory footprint. These results show that careful tokenizer--backbone--system co-design can deliver strong high-resolution generation and editing within an efficient 4B model family.

Related papers

• HiDream-O1-Image: A Natively Unified Image Generative Foundation Model with Pixel-level Unified Transformer [104.09730595701468]
  We present HiDream-O1-Image, a unified generative foundation model via pixel-space Diffusion Transformer.<n>By mapping raw image pixels, text tokens, and task-specific conditions into a single shared token space, HiDream-O1-Image achieves a structural unification of multimodal inputs.<n>Experiments show that HiDream-O1-Image excels across various generation tasks, including text-to-image generation, instruction-based editing, and subject-driven personalization.
  arXiv  Detail & Related papers  (2026-05-11T17:59:09Z)
• Nucleus-Image: Sparse MoE for Image Generation [5.769753912757775]
  We present a text-to-image generation model that exceeds leading models on GenEval, DPG-Bench, and OneIG-Bench.<n>Nucleus-Image employs a sparse mixture-of-experts (MoE) diffusion transformer architecture.<n>We construct a large-scale training corpus of 1.5B high-quality training pairs spanning 700M images.
  arXiv  Detail & Related papers  (2026-04-14T00:43:23Z)
• DeepGen 1.0: A Lightweight Unified Multimodal Model for Advancing Image Generation and Editing [67.77471070868852]
  DeepGen 1.0 is a lightweight 5B unified model for image generation and editing.<n>It is trained on only 50M samples, surpassing the 80B HunyuanImage by 28% on WISE and the 27B Qwen-Image-Edit by 37% on UniREditBench.<n>By open-sourcing our training code, weights, and datasets, we provide an efficient, high-performance alternative to democratize unified multimodal research.
  arXiv  Detail & Related papers  (2026-02-12T17:44:24Z)
• VIBE: Visual Instruction Based Editor [60.21587335143115]
  This paper presents a compact, high- throughput instruction-based image editing pipeline.<n>The pipeline is evaluated on the ImgEdit and GEdit benchmarks.<n>It generates edited images at up to 2K resolution in approximately 4 seconds on an NVIDIA H100 in BF16, without additional inference optimizations or distillation.
  arXiv  Detail & Related papers  (2026-01-05T16:17:20Z)
• E-MMDiT: Revisiting Multimodal Diffusion Transformer Design for Fast Image Synthesis under Limited Resources [12.244453688491731]
  Efficient Multimodal Diffusion Transformer (E-MMDiT) is an efficient and lightweight multimodal diffusion model with only 304M parameters for fast image synthesis.<n>Our model for 512px generation, trained with only 25M public data in 1.5 days on a single node of 8 AMD MI300X GPUs, achieves 0.66 on GenEval and easily reaches to 0.72 with some post-training techniques such as GRPO.
  arXiv  Detail & Related papers  (2025-10-31T03:13:08Z)
• Home-made Diffusion Model from Scratch to Hatch [0.9383683724544296]
  Home-made Diffusion Model (HDM) is an efficient yet powerful text-to-image diffusion model optimized for training on consumer-grade hardware.<n>HDM achieves competitive 1024x1024 generation quality while maintaining a remarkably low training cost of $535-620.
  arXiv  Detail & Related papers  (2025-09-07T14:21:57Z)
• Draw-In-Mind: Rebalancing Designer-Painter Roles in Unified Multimodal Models Benefits Image Editing [53.197392152109636]
  We introduce Draw-In-Mind (DIM), a dataset consisting of two complementary subsets: (i) DIM-T2I, containing 14M long-context image-text pairs to enhance complex instruction comprehension; and (ii) DIM-Edit, consisting of 233K chain-of-thought imaginations generated by GPT-4o, serving as explicit design blueprints for image edits.<n>DIM-4.6B-T2I/Edit achieves competitive performance on the ImgEdit and GEdit-Bench benchmarks, outperforming much larger models such as UniWorld-V1 and Step1X-Edit.
  arXiv  Detail & Related papers  (2025-09-02T06:06:52Z)
• FlowDCN: Exploring DCN-like Architectures for Fast Image Generation with Arbitrary Resolution [33.07779971446476]
  We propose FlowDCN, a purely convolution-based generative model that can efficiently generate high-quality images at arbitrary resolutions. FlowDCN achieves the state-of-the-art 4.30 sFID on $256times256$ ImageNet Benchmark and comparable resolution extrapolation results. We believe FlowDCN offers a promising solution to scalable and flexible image synthesis.
  arXiv  Detail & Related papers  (2024-10-30T02:48:50Z)
• SANA: Efficient High-Resolution Image Synthesis with Linear Diffusion Transformers [41.79064227895747]
  Sana is a text-to-image framework that can generate images up to 4096$times$4096 resolution. Sana can synthesize high-resolution, high-quality images with strong text-image alignment at a remarkably fast speed, deployable on laptop GPU.
  arXiv  Detail & Related papers  (2024-10-14T15:36:42Z)
• Latent Space Editing in Transformer-Based Flow Matching [53.75073756305241]
  Flow Matching with a transformer backbone offers the potential for scalable and high-quality generative modeling. We introduce an editing space, $u$-space, that can be manipulated in a controllable, accumulative, and composable manner. Lastly, we put forth a straightforward yet powerful method for achieving fine-grained and nuanced editing using text prompts.
  arXiv  Detail & Related papers  (2023-12-17T21:49:59Z)
• Dual-former: Hybrid Self-attention Transformer for Efficient Image Restoration [6.611849560359801]
  We present Dual-former, which combines the powerful global modeling ability of self-attention modules and the local modeling ability of convolutions in an overall architecture. Experiments demonstrate that Dual-former achieves a 1.91dB gain over the state-of-the-art MAXIM method on the Indoor dataset for single image dehazing. For single image deraining, it exceeds the SOTA method by 0.1dB PSNR on the average results of five datasets with only 21.5% GFLOPs.
  arXiv  Detail & Related papers  (2022-10-03T16:39:21Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.