<h1 id="qwen3-vl">Qwen3-VL</h1>Meet Qwen3-VL — the most powerful vision-language model in the Qwen series to date.This generation delivers comprehensive upgrades across the board: superior text understanding &amp; generation, deeper visual perception &amp; reasoning, extended context length, enhanced spatial and video dynamics comprehension, and stronger agent interaction capabilities.Available in Dense and MoE architectures that scale from edge to cloud, with Instruct and reasoning‑enhanced Thinking editions for flexible, on‑demand deployment.Key Enhancements:<ul><li>Visual Agent: Operates PC/mobile GUIs—recognizes elements, understands functions, invokes tools, completes tasks.</li><li>Visual Coding Boost: Generates Draw.io/HTML/CSS/JS from images/videos.</li><li>Advanced Spatial Perception: Judges object positions, viewpoints, and occlusions; provides stronger 2D grounding and enables 3D grounding for spatial reasoning and embodied AI.</li><li>Long Context &amp; Video Understanding: Native 256K context, expandable to 1M; handles books and hours-long video with full recall and second-level indexing.</li><li>Enhanced Multimodal Reasoning: Excels in STEM/Math—causal analysis and logical, evidence-based answers.</li><li>Upgraded Visual Recognition: Broader, higher-quality pretraining is able to “recognize everything”—celebrities, anime, products, landmarks, flora/fauna, etc.</li><li>Expanded OCR: Supports 32 languages (up from 19); robust in low light, blur, and tilt; better with rare/ancient characters and jargon; improved long-document structure parsing.</li><li>Text Understanding on par with pure LLMs: Seamless text–vision fusion for lossless, unified comprehension.</li></ul>Model Architecture Updates:<img src="https://qianwen-res.oss-accelerate.aliyuncs.com/Qwen3-VL/qwen3vl_arc.jpg" width="80%" /><ol><li>Interleaved-MRoPE: Full‑frequency allocation over time, width, and height via robust positional embeddings, enhancing long‑horizon video reasoning.</li><li>DeepStack: Fuses multi‑level ViT features to capture fine‑grained details and sharpen image–text alignment.</li><li>Text–Timestamp Alignment: Moves beyond T‑RoPE to precise, timestamp‑grounded event localization for stronger video temporal modeling.</li></ol>

Qwen3-VL

qwen3VL_8b.safetensors

Boogu-Image Turbo is the speed-distilled text-to-image build in the Boogu-Image 0.1 family. It carries the same 10B-parameter backbone as <a href="/models/2714299" rel="ugc">Boogu-Image Base</a> but is tuned to produce photorealistic results in just 4 steps with CFG 0, trading some of Base's diversity for a large jump in generation speed.
Originally released by the <a href="https://huggingface.co/Boogu" rel="ugc">Boogu team</a> on Hugging Face. All credit for the model goes to the Boogu project and its contributors. Civitai is hosting a mirror so creators can run it on-site without a local setup - please head to the <a href="https://github.com/boogu-project/Boogu-Image" rel="ugc">original repo</a> and <a href="https://boogu.org" rel="ugc">project site</a> for weights, updates, and to support the project directly.
<h3>Built by the Boogu project</h3>
<ul>
<li><a href="https://boogu.org" rel="ugc">Boogu project</a> - model training and release</li>
<li><a href="https://huggingface.co/Boogu" rel="ugc">Hugging Face org</a> and <a href="https://modelscope.cn/organization/Boogu" rel="ugc">ModelScope org</a> - weights and variant downloads</li>
<li>Acknowledged upstream influences: Qwen-Image, Z-Image, OmniGen2, FLUX, and DeepSeek</li>
</ul>
<h3>What it does</h3>
Turbo is the build to reach for when throughput matters: fast iteration, batch work, and interactive generation. It keeps the family's strengths in photography, bilingual Chinese and English text rendering, stylization, and poster design, at a fraction of the step count. Recommended starting point is 4 steps, CFG 0.
<h3>Versions on Civitai</h3>
We mirror the v0.1 release in three diffusion builds: <code>bf16</code> full precision (~20GB), <code>fp8_scaled</code> (~10GB), and <code>nvfp4</code> (~5.7GB). The required Qwen3-VL 8B text encoder and the Flux VAE are bundled. Note: this is a full distilled checkpoint, not a LoRA.
<h3>Related Boogu models</h3>
<a href="/models/2714299" rel="ugc">Boogu-Image Base</a> is the full-quality foundation model at 25 to 50 steps. <a href="/models/2714541" rel="ugc">Boogu-Image Edit</a> adds image-to-image editing.
<h3>Known limitations</h3>
The Boogu team is candid that 0.1 is an early research release: limited world knowledge for brands, celebrities, and landmarks; instability in strict subject preservation; occasional text typos or layout drift; and artifacts in complex poses, small faces, and limbs.
<h3>Links</h3>
<ul>
<li>Source: <a href="https://huggingface.co/Boogu" rel="ugc">Hugging Face</a> | <a href="https://github.com/boogu-project/Boogu-Image" rel="ugc">GitHub</a> | <a href="https://modelscope.cn/organization/Boogu" rel="ugc">ModelScope</a></li>
<li>Project site: <a href="https://boogu.org" rel="ugc">boogu.org</a></li>
<li>Gallery: <a href="https://boogu-gallery.netlify.app/" rel="ugc">boogu-gallery.netlify.app</a></li>
</ul>
Boogu-Image 0.1 is a research project, not an official product release. The Boogu team asks that you not build unauthorized commercial products under the Boogu name. Licensed under Apache 2.0.

Boogu-Image Turbo

flux1_vae_bf16.safetensors

booguImageTurbo_v01_txt.safetensors

booguImageTurbo_v01_bf16.safetensors

booguImageTurbo_v01_nf4.safetensors

booguImageTurbo_v01_fp8.safetensors

v0.1

Boogu-Image Edit is the image editing build in the Boogu-Image 0.1 family, a 10B-parameter model that takes a text prompt plus a source image and returns an edited result. It handles object insertion, replacement, and removal, attribute changes, and text editing while keeping the rest of the scene intact.
Originally released by the <a href="https://huggingface.co/Boogu" rel="ugc">Boogu team</a> on Hugging Face. All credit for the model goes to the Boogu project and its contributors. Civitai is hosting a mirror so creators can run it on-site without a local setup - please head to the <a href="https://github.com/boogu-project/Boogu-Image" rel="ugc">original repo</a> and <a href="https://boogu.org" rel="ugc">project site</a> for weights, updates, and to support the project directly.
<h3>Built by the Boogu project</h3>
<ul>
<li><a href="https://boogu.org" rel="ugc">Boogu project</a> - model training and release</li>
<li><a href="https://huggingface.co/Boogu" rel="ugc">Hugging Face org</a> and <a href="https://modelscope.cn/organization/Boogu" rel="ugc">ModelScope org</a> - weights and variant downloads</li>
<li>Acknowledged upstream influences: Qwen-Image, Z-Image, OmniGen2, FLUX, and DeepSeek</li>
</ul>
<h3>What it does</h3>
Edit is a text-plus-image-to-image model for targeted changes: add or remove objects, swap attributes (color, material, style), and edit rendered text, all driven by a natural-language instruction. It shares the family's bilingual Chinese and English text handling and runs at 25 to 50 steps with CFG in the 2.0 to 5.0 range.
<h3>Versions on Civitai</h3>
We mirror the v0.1 release in three diffusion builds: <code>bf16</code> full precision (~20GB), <code>fp8_scaled</code> (~10GB), and <code>nvfp4</code> (~5.7GB). The Qwen3-VL 8B text encoder and the Flux VAE are bundled so you can assemble a complete editing workflow.
<h3>Related Boogu models</h3>
<a href="/models/2714299" rel="ugc">Boogu-Image Base</a> is the text-to-image foundation model. <a href="/models/2714686" rel="ugc">Boogu-Image Turbo</a> is the fast 4-step text-to-image variant.
<h3>Known limitations</h3>
The Boogu team is candid that 0.1 is an early research release: limited world knowledge for brands, celebrities, and landmarks; instability in strict subject preservation; occasional text typos or layout drift; and artifacts in complex poses, small faces, and limbs.
<h3>Links</h3>
<ul>
<li>Source: <a href="https://huggingface.co/Boogu" rel="ugc">Hugging Face</a> | <a href="https://github.com/boogu-project/Boogu-Image" rel="ugc">GitHub</a> | <a href="https://modelscope.cn/organization/Boogu" rel="ugc">ModelScope</a></li>
<li>Project site: <a href="https://boogu.org" rel="ugc">boogu.org</a></li>
<li>Gallery: <a href="https://boogu-gallery.netlify.app/" rel="ugc">boogu-gallery.netlify.app</a></li>
</ul>
Boogu-Image 0.1 is a research project, not an official product release. The Boogu team asks that you not build unauthorized commercial products under the Boogu name. Licensed under Apache 2.0.

Boogu-Image Edit

booguImageEdit_v01_fp8.safetensors

booguImageEdit_v01_bf16.safetensors

booguImageEdit_v01_txt.safetensors

booguImageEdit_v01_nf4.safetensors

Boogu-Image Edit is the image editing build in the Boogu-Image 0.1 family, a 10B-parameter model that takes a text prompt plus a source image and returns an edited result. It handles object insertion, replacement, and removal, attribute changes, and text editing while keeping the rest of the scene intact.Originally released by the <a target="_blank" rel="ugc" href="https://huggingface.co/Boogu">Boogu team</a> on Hugging Face. All credit for the model goes to the Boogu project and its contributors. Civitai is hosting a mirror so creators can run it on-site without a local setup - please head to the <a target="_blank" rel="ugc" href="https://github.com/boogu-project/Boogu-Image">original repo</a> and <a target="_blank" rel="ugc" href="https://boogu.org">project site</a> for weights, updates, and to support the project directly.<h3 id="built-by-the-boogu-project">Built by the Boogu project</h3><ul><li><a target="_blank" rel="ugc" href="https://boogu.org">Boogu project</a> - model training and release</li><li><a target="_blank" rel="ugc" href="https://huggingface.co/Boogu">Hugging Face org</a> and <a target="_blank" rel="ugc" href="https://modelscope.cn/organization/Boogu">ModelScope org</a> - weights and variant downloads</li><li>Acknowledged upstream influences: Qwen-Image, Z-Image, OmniGen2, FLUX, and DeepSeek</li></ul><h3 id="what-it-does">What it does</h3>Edit is a text-plus-image-to-image model for targeted changes: add or remove objects, swap attributes (color, material, style), and edit rendered text, all driven by a natural-language instruction. It shares the family's bilingual Chinese and English text handling and runs at 25 to 50 steps with CFG in the 2.0 to 5.0 range.<h3 id="versions-on-civitai">Versions on Civitai</h3>We mirror the v0.1 release in three diffusion builds: <code>bf16</code> full precision (~20GB), <code>fp8_scaled</code> (~10GB), and <code>nvfp4</code> (~5.7GB). The Qwen3-VL 8B text encoder and the Flux VAE are bundled so you can assemble a complete editing workflow.<h3 id="related-boogu-models">Related Boogu models</h3><a target="_blank" rel="ugc" href="/models/2714299">Boogu-Image Base</a> is the text-to-image foundation model. <a target="_blank" rel="ugc" href="/models/2714686">Boogu-Image Turbo</a> is the fast 4-step text-to-image variant.<h3 id="known-limitations">Known limitations</h3>The Boogu team is candid that 0.1 is an early research release: limited world knowledge for brands, celebrities, and landmarks; instability in strict subject preservation; occasional text typos or layout drift; and artifacts in complex poses, small faces, and limbs.<h3 id="links">Links</h3><ul><li>Source: <a target="_blank" rel="ugc" href="https://huggingface.co/Boogu">Hugging Face</a> | <a target="_blank" rel="ugc" href="https://github.com/boogu-project/Boogu-Image">GitHub</a> | <a target="_blank" rel="ugc" href="https://modelscope.cn/organization/Boogu">ModelScope</a></li><li>Project site: <a target="_blank" rel="ugc" href="https://boogu.org">boogu.org</a></li><li>Gallery: <a target="_blank" rel="ugc" href="https://boogu-gallery.netlify.app/">boogu-gallery.netlify.app</a></li></ul>Boogu-Image 0.1 is a research project, not an official product release. The Boogu team asks that you not build unauthorized commercial products under the Boogu name. Licensed under Apache 2.0.

Boogu-Image Edit-turbo

booguImageEditTurbo_v01_fp8.safetensors

booguImageEditTurbo_v01_bf16.safetensors

booguImageEditTurbo_v01_txt.safetensors

Ideogram 4 is a foundation model trained entirely from scratch, not a fine-tune or distillation of any existing checkpoint. It is a flow-matching text-to-image model built on a fully single-stream Diffusion Transformer (DiT) architecture.Architecture:<ul><li>Fully single-stream DiT. Text and image tokens are concatenated into one unified sequence and processed through the same 34-layer transformer, with no separate text or image branches. This enables deep cross-modal interaction at every layer.</li><li>Vision-language model as text encoder. Instead of a text-only encoder like CLIP or T5, Ideogram 4 uses <a target="_blank" rel="ugc" href="https://huggingface.co/Qwen/Qwen3-VL-8B-Instruct">Qwen3-VL-8B-Instruct</a>, a full vision-language model that provides far richer understanding of visual concepts. Hidden states are extracted from 13 intermediate layers and concatenated, giving the model multi-scale semantic features ranging from surface-level token information to deep compositional understanding.</li><li>Dual-branch classifier-free guidance. The conditional (positive) and unconditional (negative) branches can be independently refined, enabling separate control over prompt adherence and image quality.</li><li>Flexible resolution. Native support for any resolution from 256 to 2048 (multiples of 16), with aspect ratios up to 6:1. A single model handles everything from square thumbnails to ultrawide banners, with the noise schedule auto-adjusting per resolution.</li></ul>Key Capabilities:<ul><li>Extreme controllability. Ideogram 4 is trained on structured JSON captions, giving users unprecedented control over composition, style, lighting, color palette, typography, and spatial layout, all from a single prompt.</li><li>State-of-the-art text rendering. Ideogram 4 delivers best-in-class in-image text generation (signage, logos, captions, watermarks, multi-line text) with high fidelity directly from the prompt.</li><li>Spatial layout control. Bounding-box coordinates in the prompt allow explicit placement of subjects, text elements, and background regions.</li><li>Color palette conditioning. Specify hex colors in the prompt to steer the image's dominant color scheme.</li></ul>For full architecture details, see <a target="_blank" rel="ugc" href="https://github.com/ideogram-oss/ideogram4/blob/main/docs/model_architecture.md">docs/model_architecture.md</a>. For a walkthrough of how the pipeline components fit together, see <a target="_blank" rel="ugc" href="https://github.com/ideogram-oss/ideogram4/blob/main/docs/pipeline.md">docs/pipeline.md</a>.<h2 id="prompting-guide">Prompting Guide</h2>Ideogram 4 is trained exclusively on structured JSON captions. While plain-text prompts work, you will get the best results by providing a JSON object that follows our caption schema.Key points:<ul><li>Use JSON prompts for maximum controllability — the model was trained on them and understands the structure natively.</li><li>Color palette conditioning — specify a <code>colour_palette</code> array of hex colors in the style description to steer the image's color scheme.</li><li>Aspect ratio flexibility — Ideogram 4 supports a wide range of aspect ratios (any multiple-of-16 resolution from 256 to 2048 on each side). This is a key advantage for practical use: portraits, landscapes, banners, phone wallpapers, social media formats, etc.</li><li>Bounding-box layout — specify <code>bbox</code> coordinates in the prompt to explicitly place subjects, text elements, and background regions.</li><li>Compositional control — use <code>compositional_deconstruction</code> with bounding boxes and per-element descriptions for precise spatial layout.</li></ul>Why JSON-only training? We train exclusively on JSON so that training and inference share a single, common prompt format. The training captions themselves are deliberately extremely descriptive: each JSON exhaustively describes everything in the image to maximize training efficiency. The more text-to-image relationships each caption pins down, the more grounded supervision the model extracts from a single training pair, rather than having to infer those relationships across many sparsely-captioned samples.Why JSON at inference time? Because the model was trained on captions that name every object explicitly, the most reliable way to get every requested object rendered is to mirror that pattern. Plain-text prompts still work, but won't perform as well since the model was only trained on structured JSON captions.Don't want to write JSON by hand? That's what magic prompt is for: it uses an LLM to expand a plain-text prompt into a full structured caption before generation, so you get JSON-quality results from a casual prompt. It runs by default in <code>run_inference.py</code> (see the <a target="_blank" rel="ugc" href="https://huggingface.co/ideogram-ai/ideogram-4-fp8#cli">CLI</a> section).See <a target="_blank" rel="ugc" href="https://github.com/ideogram-oss/ideogram4/blob/main/docs/prompting.md">docs/prompting.md</a> for a full guide.

Ideogram 4

ideogram4_fp32.safetensors

ideogram4_3129012.safetensors

ideogram4_3129014.safetensors

ideogram4_txt_3128943.safetensors

ideogram4_3128958.safetensors

ideogram4_3128980.safetensors

ideogram4_txt_3128946.safetensors

ideogram4_3128969.safetensors

ideogram4_3129032.safetensors

ideogram-4

Boogu-Image Base is the foundation text-to-image model in the Boogu-Image 0.1 family, a 10B-parameter multimodal system built for photographic realism, dense bilingual text rendering, and broad stylistic range. It is the model to start with when you want maximum diversity and prompt adherence, and the parent that the Turbo and Edit builds derive from.
Originally released by the <a href="https://huggingface.co/Boogu" rel="ugc">Boogu team</a> on Hugging Face. All credit for the model goes to the Boogu project and its contributors. Civitai is hosting a mirror so creators can run it on-site without a local setup - please head to the <a href="https://github.com/boogu-project/Boogu-Image" rel="ugc">original repo</a> and <a href="https://boogu.org" rel="ugc">project site</a> for weights, updates, and to support the project directly.
<h3>Built by the Boogu project</h3>
<ul>
<li><a href="https://boogu.org" rel="ugc">Boogu project</a> - model training and release</li>
<li><a href="https://huggingface.co/Boogu" rel="ugc">Hugging Face org</a> and <a href="https://modelscope.cn/organization/Boogu" rel="ugc">ModelScope org</a> - weights and variant downloads</li>
<li>Acknowledged upstream influences: Qwen-Image, Z-Image, OmniGen2, FLUX, and DeepSeek</li>
</ul>
<h3>What it does</h3>
Base targets photography-grade output with natural lighting and composition, bilingual Chinese and English text rendering across complex layouts, stylization across many artistic genres, and poster and product design. It runs at 25 to 50 steps with CFG in the 2.0 to 5.0 range.
<h3>Versions on Civitai</h3>
We mirror the v0.1 release in three diffusion builds: <code>bf16</code> full precision (~20GB), <code>fp8_scaled</code> (~10GB), and <code>nvfp4</code> (~5.7GB). The required Qwen3-VL 8B text encoder and the Flux VAE are bundled so you can assemble a complete workflow.
<h3>Related Boogu models</h3>
<a href="/models/2714686" rel="ugc">Boogu-Image Turbo</a> is the distilled 4-step variant for fast generation. <a href="/models/2714541" rel="ugc">Boogu-Image Edit</a> adds image-to-image editing.
<h3>Known limitations</h3>
The Boogu team is candid that 0.1 is an early research release: limited world knowledge for brands, celebrities, and landmarks; instability in strict subject preservation; occasional text typos or layout drift; and artifacts in complex poses, small faces, and limbs.
<h3>Links</h3>
<ul>
<li>Source: <a href="https://huggingface.co/Boogu" rel="ugc">Hugging Face</a> | <a href="https://github.com/boogu-project/Boogu-Image" rel="ugc">GitHub</a> | <a href="https://modelscope.cn/organization/Boogu" rel="ugc">ModelScope</a></li>
<li>Project site: <a href="https://boogu.org" rel="ugc">boogu.org</a></li>
<li>Gallery: <a href="https://boogu-gallery.netlify.app/" rel="ugc">boogu-gallery.netlify.app</a></li>
</ul>
Boogu-Image 0.1 is a research project, not an official product release. The Boogu team asks that you not build unauthorized commercial products under the Boogu name. Licensed under Apache 2.0.

Files with SHA256: 4ba424cf62e51392e4d1a39933e803706f4e823c1065f36aaf149c6453f66bcd