{"id":15059,"date":"2026-05-16T10:28:02","date_gmt":"2026-05-16T09:28:02","guid":{"rendered":"https:\/\/www.keris-studio.fr\/blog\/?p=15059"},"modified":"2026-05-16T12:07:08","modified_gmt":"2026-05-16T11:07:08","slug":"ai-3d-tests","status":"publish","type":"post","link":"https:\/\/www.keris-studio.fr\/blog\/?p=15059","title":{"rendered":"Ai 3D tests"},"content":{"rendered":"<p>Starting base<\/p>\n<p>This image serves as the foundation from which the AI extrapolates the three-dimensional geometry (mesh), vertices, and topology of the final 3D model. <br \/><img loading=\"lazy\" decoding=\"async\" width=\"1940\" height=\"816\" class=\"wp-image-15060\" src=\"https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-1.png\" srcset=\"https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-1.png 1940w, https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-1-300x126.png 300w, https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-1-1024x431.png 1024w, https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-1-768x323.png 768w, https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-1-1536x646.png 1536w\" sizes=\"auto, (max-width: 1940px) 100vw, 1940px\" \/><!--more--><\/p>\n<h1><a href=\"https:\/\/studio.tripo3d.ai\/\">tripo3d<\/a><\/h1>\n<p><a href=\"https:\/\/studio.tripo3d.ai\/\">https:\/\/studio.tripo3d.ai\/<\/a><\/p>\n<p>Tripo3D is an AI-powered service designed to generate 3D models, primarily from input sources like 2D images or text prompts. It is a tool used for \u00ab\u00a0Ai 3D tests\u00a0\u00bb and specifically listed as an \u00ab\u00a0Image -&gt; 3D\u00a0\u00bb solution. The kind of models generated are standard polygonal geometry (meshes), characterized by specific counts of faces, vertices, and a triangle topology. The platform integrates several features to create and refine these models.<\/p>\n<ul>\n<li><strong>Generation and Texturing:<\/strong> It allows for model creation using text or image input, and includes specialized features for texturing, such as \u00ab\u00a0Texture Gen,\u00a0\u00bb \u00ab\u00a0HD Texture,\u00a0\u00bb and a \u00ab\u00a0PBR Generator\u00a0\u00bb.<\/li>\n<li><strong>Refinement:<\/strong> The platform offers post-processing capabilities like \u00ab\u00a0Retopology\u00a0\u00bb and \u00ab\u00a0Segmentation\u00a0\u00bb to clean up and optimize the generated 3D meshes.<\/li>\n<\/ul>\n<p><img loading=\"lazy\" decoding=\"async\" width=\"2048\" height=\"981\" class=\"wp-image-15061\" src=\"https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-2.png\" srcset=\"https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-2.png 2048w, https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-2-300x144.png 300w, https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-2-1024x491.png 1024w, https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-2-768x368.png 768w, https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-2-1536x736.png 1536w\" sizes=\"auto, (max-width: 2048px) 100vw, 2048px\" \/><\/p>\n<p><img loading=\"lazy\" decoding=\"async\" width=\"1342\" height=\"726\" class=\"wp-image-15062\" src=\"https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-3.png\" srcset=\"https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-3.png 1342w, https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-3-300x162.png 300w, https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-3-1024x554.png 1024w, https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-3-768x415.png 768w\" sizes=\"auto, (max-width: 1342px) 100vw, 1342px\" \/> <img loading=\"lazy\" decoding=\"async\" width=\"1522\" height=\"733\" class=\"wp-image-15063\" src=\"https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-4.png\" srcset=\"https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-4.png 1522w, https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-4-300x144.png 300w, https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-4-1024x493.png 1024w, https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-4-768x370.png 768w\" sizes=\"auto, (max-width: 1522px) 100vw, 1522px\" \/> <img loading=\"lazy\" decoding=\"async\" width=\"1362\" height=\"740\" class=\"wp-image-15064\" src=\"https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-5.png\" srcset=\"https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-5.png 1362w, https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-5-300x163.png 300w, https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-5-1024x556.png 1024w, https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-5-768x417.png 768w\" sizes=\"auto, (max-width: 1362px) 100vw, 1362px\" \/> <img loading=\"lazy\" decoding=\"async\" width=\"1425\" height=\"709\" class=\"wp-image-15065\" src=\"https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-6.png\" srcset=\"https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-6.png 1425w, https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-6-300x149.png 300w, https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-6-1024x509.png 1024w, https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-6-768x382.png 768w\" sizes=\"auto, (max-width: 1425px) 100vw, 1425px\" \/><\/p>\n<p><strong>Examin 3D<br \/><\/strong><strong><a href=\"https:\/\/studio.tripo3d.ai\/3d-model\/8a311ead-ef9d-4e40-8591-0d92e66774ee?invite_code=87ON1J\">https:\/\/studio.tripo3d.ai\/3d-model\/8a311ead-ef9d-4e40-8591-0d92e66774ee?invite_code=87ON1J<\/a><\/strong><\/p>\n<h1><a href=\"https:\/\/huggingface.co\/spaces\/TencentARC\/Pixal3D\">Pixal3D<\/a><\/h1>\n<p>Pixal3D is a new, high-fidelity <strong>pixel-aligned 3D generation paradigm<\/strong> designed to create 3D assets from 2D images.Where it comes from<a href=\"https:\/\/docs.google.com\/document\/d\/1YwtuexRo1HiyYnMiWBvKHLIg6si8KJ8EuXcMw8d5sFM\/edit\">1<\/a><\/p>\n<p>The model is associated with <strong>TencentARC<\/strong>, as seen in the URL for the Hugging Face space provided in your document.Why it is different<a href=\"https:\/\/docs.google.com\/document\/d\/1YwtuexRo1HiyYnMiWBvKHLIg6si8KJ8EuXcMw8d5sFM\/edit\">1<\/a><\/p>\n<p>Pixal3D addresses a major challenge in image-to-3D generation: achieving pixel-level faithfulness to the input image.<\/p>\n<p>Previous 3D-native generators often synthesize shapes in canonical space and inject image cues via attention, which can leave pixel-to-3D associations ambiguous.<a href=\"https:\/\/docs.google.com\/document\/d\/1YwtuexRo1HiyYnMiWBvKHLIg6si8KJ8EuXcMw8d5sFM\/edit\">1<\/a><\/p>\n<p>Pixal3D differs by:<\/p>\n<ul>\n<li><strong>Pixel-Aligned Generation:<\/strong> It generates the 3D model directly in a way that is consistent with the input view, instead of synthesizing shapes in a canonical space.<\/li>\n<li><strong>Direct Correspondence:<\/strong> It introduces a <strong>pixel back-projection conditioning scheme<\/strong> that explicitly lifts multi-scale image features into a 3D feature volume, establishing a direct and unambiguous pixel-to-3D correspondence.<\/li>\n<li><strong>Scalability and Fidelity:<\/strong> The approach is highly scalable and substantially improves fidelity, approaching the level of traditional 3D reconstruction methods.<\/li>\n<li><strong>Versatility:<\/strong> The pixel-aligned paradigm can naturally extend to multi-view generation and benefits scene synthesis by allowing for a modular pipeline that produces object-separated 3D scenes from images.<\/li>\n<\/ul>\n<p>Recent advancements in 3D generative modeling have significantly improved the generation realism, yet the field is still hampered by existing representations, which struggle to capture assets with complex topologies and detailed appearance. This paper presents an approach for learning a structured latent representation from native 3D data to address this challenge. At its core is a new sparse voxel structure called O-Voxel, an omni-voxel representation that encodes both geometry and appearance. O-Voxel can robustly model arbitrary topology, including open, non-manifold, and fully-enclosed surfaces, while capturing comprehensive surface attributes beyond texture color, such as physically-based rendering parameters. Based on O-Voxel, we design a Sparse Compression VAE which provides a high spatial compression rate and a compact latent space. We train large-scale flow-matching models comprising 4B parameters for 3D generation using diverse public 3D asset datasets. Despite their scale, inference remains highly efficient. Meanwhile, the geometry and material quality of our generated assets far exceed those of existing models. We believe our approach offers a significant advancement in 3D generative modeling.<\/p>\n<p>source : <a href=\"https:\/\/arxiv.org\/html\/2512.14692v1\">https:\/\/arxiv.org\/html\/2512.14692v1<\/a><\/p>\n<p><a href=\"https:\/\/huggingface.co\/spaces\/TencentARC\/Pixal3D\">https:\/\/huggingface.co\/spaces\/TencentARC\/Pixal3D<\/a><\/p>\n<p><a href=\"https:\/\/ldyang694.github.io\/projects\/pixal3d\/\">https:\/\/ldyang694.github.io\/projects\/pixal3d\/<\/a><\/p>\n<p><br \/><img loading=\"lazy\" decoding=\"async\" width=\"2048\" height=\"744\" class=\"wp-image-15066\" src=\"https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-7.png\" srcset=\"https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-7.png 2048w, https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-7-300x109.png 300w, https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-7-1024x372.png 1024w, https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-7-768x279.png 768w, https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-7-1536x558.png 1536w\" sizes=\"auto, (max-width: 2048px) 100vw, 2048px\" \/><\/p>\n<p><img loading=\"lazy\" decoding=\"async\" width=\"783\" height=\"397\" class=\"wp-image-15067\" src=\"https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-8.png\" srcset=\"https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-8.png 783w, https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-8-300x152.png 300w, https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-8-768x389.png 768w\" sizes=\"auto, (max-width: 783px) 100vw, 783px\" \/> <img loading=\"lazy\" decoding=\"async\" width=\"944\" height=\"457\" class=\"wp-image-15068\" src=\"https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-9.png\" srcset=\"https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-9.png 944w, https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-9-300x145.png 300w, https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-9-768x372.png 768w\" sizes=\"auto, (max-width: 944px) 100vw, 944px\" \/><\/p>\n<p><img loading=\"lazy\" decoding=\"async\" width=\"1438\" height=\"828\" class=\"wp-image-15069\" src=\"https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-10.png\" srcset=\"https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-10.png 1438w, https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-10-300x173.png 300w, https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-10-1024x590.png 1024w, https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-10-768x442.png 768w\" sizes=\"auto, (max-width: 1438px) 100vw, 1438px\" \/><\/p>\n<p><img loading=\"lazy\" decoding=\"async\" width=\"636\" height=\"290\" class=\"wp-image-15070\" src=\"https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-11.png\" srcset=\"https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-11.png 636w, https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-11-300x137.png 300w\" sizes=\"auto, (max-width: 636px) 100vw, 636px\" \/> <img loading=\"lazy\" decoding=\"async\" width=\"624\" height=\"311\" class=\"wp-image-15071\" src=\"https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-12.png\" srcset=\"https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-12.png 624w, https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-12-300x150.png 300w\" sizes=\"auto, (max-width: 624px) 100vw, 624px\" \/><\/p>\n<p><a href=\"https:\/\/fal.ai\/models\/fal-ai\/pixal3d?share=5bdeafee-ed37-4256-948e-7e53626f9d6d\"><strong>Examin 3D<\/strong><br \/>https:\/\/fal.ai\/models\/fal-ai\/pixal3d?share=5bdeafee-ed37-4256-948e-7e53626f9d6d<\/a><\/p>\n<h1><a href=\"https:\/\/www.meshy.ai\/fr\/workspace?model-tab=image-to-3d\">meshy.ai<\/a><\/h1>\n<p><a href=\"https:\/\/www.meshy.ai\/fr\/workspace?model-tab=image-to-3d\">https:\/\/www.meshy.ai\/fr\/workspace?model-tab=image-to-3d<\/a><\/p>\n<p>Meshy.ai is a platform committed to <strong>simplifying and accelerating the 3D content creation process<\/strong>. It provides intuitive tools aimed at empowering artists, designers, and enthusiasts to bring their imagination to life.<a href=\"https:\/\/mail.google.com\/mail\/u\/0\/#all\/18bf2f00f7e29ecc\">1<\/a><\/p>\n<p>The Meshy workspace, based on the URL provided, is specifically focused on the <strong>Image-to-3D<\/strong> model tab, placing it within the category of AI tools that generate 3D assets from a 2D input.<\/p>\n<p><img loading=\"lazy\" decoding=\"async\" width=\"1876\" height=\"1168\" class=\"wp-image-15072\" src=\"https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-13.png\" srcset=\"https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-13.png 1876w, https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-13-300x187.png 300w, https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-13-1024x638.png 1024w, https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-13-768x478.png 768w, https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-13-1536x956.png 1536w\" sizes=\"auto, (max-width: 1876px) 100vw, 1876px\" \/> <img loading=\"lazy\" decoding=\"async\" width=\"1275\" height=\"830\" class=\"wp-image-15073\" src=\"https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-14.png\" srcset=\"https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-14.png 1275w, https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-14-300x195.png 300w, https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-14-1024x667.png 1024w, https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-14-768x500.png 768w\" sizes=\"auto, (max-width: 1275px) 100vw, 1275px\" \/> <img loading=\"lazy\" decoding=\"async\" width=\"1346\" height=\"643\" class=\"wp-image-15074\" src=\"https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-15.png\" srcset=\"https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-15.png 1346w, https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-15-300x143.png 300w, https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-15-1024x489.png 1024w, https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-15-768x367.png 768w\" sizes=\"auto, (max-width: 1346px) 100vw, 1346px\" \/><\/p>\n<p><img loading=\"lazy\" decoding=\"async\" width=\"1390\" height=\"687\" class=\"wp-image-15075\" src=\"https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-16.png\" srcset=\"https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-16.png 1390w, https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-16-300x148.png 300w, https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-16-1024x506.png 1024w, https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-16-768x380.png 768w\" sizes=\"auto, (max-width: 1390px) 100vw, 1390px\" \/> <img loading=\"lazy\" decoding=\"async\" width=\"1425\" height=\"654\" class=\"wp-image-15076\" src=\"https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-17.png\" srcset=\"https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-17.png 1425w, https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-17-300x138.png 300w, https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-17-1024x470.png 1024w, https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-17-768x352.png 768w\" sizes=\"auto, (max-width: 1425px) 100vw, 1425px\" \/> <img loading=\"lazy\" decoding=\"async\" width=\"1375\" height=\"671\" class=\"wp-image-15078\" src=\"https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-18-1.png\" srcset=\"https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-18-1.png 1375w, https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-18-1-300x146.png 300w, https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-18-1-1024x500.png 1024w, https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-18-1-768x375.png 768w\" sizes=\"auto, (max-width: 1375px) 100vw, 1375px\" \/> <img loading=\"lazy\" decoding=\"async\" width=\"1454\" height=\"637\" class=\"wp-image-15079\" src=\"https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-19.png\" srcset=\"https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-19.png 1454w, https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-19-300x131.png 300w, https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-19-1024x449.png 1024w, https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-19-768x336.png 768w\" sizes=\"auto, (max-width: 1454px) 100vw, 1454px\" \/><\/p>\n<p><a href=\"https:\/\/www.meshy.ai\/s\/gWMXUE\"><strong>Examin 3D<\/strong><br \/>https:\/\/www.meshy.ai\/s\/gWMXUE<\/a><\/p>\n<h1>hitem3d<\/h1>\n<p><a href=\"https:\/\/www.hitem3d.ai\/creative-space\">https:\/\/www.hitem3d.ai\/creative-space<\/a><\/p>\n<p><strong>Creative Space &#8211; Hi3D<\/strong> is an all-in-one AI image processing and 3D model generation platform. It is used to convert standard flat images or multi-view photos into 3D assets.<\/p>\n<p>Key features and options of the platform include:<\/p>\n<ul>\n<li><strong>Input Requirements:<\/strong> Users must upload a mandatory front-view image (under 20MB) to start the process, with additional side and back views remaining optional.<\/li>\n<li><strong>Model Selection:<\/strong> Users can choose from various AI model versions (such as v1.5, v2.0, and v2.1) to control the desired texture realism and structural stability of the output.<\/li>\n<li><strong>Processing Modes:<\/strong> Processing speed can be toggled between a \u00ab\u00a0Fast\u00a0\u00bb mode (about two minutes) and a \u00ab\u00a0Quality\u00a0\u00bb mode optimized for production-ready models.<\/li>\n<li><strong>Generation Output:<\/strong> Creators can choose to generate just the underlying geometry or a combination of both geometry and texture.<\/li>\n<li><strong>Specialized Features:<\/strong> It includes a \u00ab\u00a0Portrait mode\u00a0\u00bb to refine stray hairs, clean up backgrounds, and restore missing details for sharper generations.<\/li>\n<li><strong>Advanced Settings:<\/strong> Users can adjust specific parameters like poly count, alignment, and UV unwrapping.<\/li>\n<li><strong>Sharing and Access:<\/strong> Generated assets can be shared publicly under a CC BY 4.0 license, while private generation requires an upgraded account. A free plan is offered, but upgrading provides extra monthly credits.<\/li>\n<\/ul>\n<p><img loading=\"lazy\" decoding=\"async\" width=\"2048\" height=\"926\" class=\"wp-image-15080\" src=\"https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-20.png\" srcset=\"https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-20.png 2048w, https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-20-300x136.png 300w, https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-20-1024x463.png 1024w, https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-20-768x347.png 768w, https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-20-1536x695.png 1536w\" sizes=\"auto, (max-width: 2048px) 100vw, 2048px\" \/><\/p>\n<p><img loading=\"lazy\" decoding=\"async\" width=\"1644\" height=\"861\" class=\"wp-image-15081\" src=\"https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-21.png\" srcset=\"https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-21.png 1644w, https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-21-300x157.png 300w, https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-21-1024x536.png 1024w, https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-21-768x402.png 768w, https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-21-1536x804.png 1536w\" sizes=\"auto, (max-width: 1644px) 100vw, 1644px\" \/><\/p>\n<p><img loading=\"lazy\" decoding=\"async\" width=\"1658\" height=\"877\" class=\"wp-image-15082\" src=\"https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-22.png\" srcset=\"https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-22.png 1658w, https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-22-300x159.png 300w, https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-22-1024x542.png 1024w, https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-22-768x406.png 768w, https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-22-1536x812.png 1536w\" sizes=\"auto, (max-width: 1658px) 100vw, 1658px\" \/> <img loading=\"lazy\" decoding=\"async\" width=\"1558\" height=\"792\" class=\"wp-image-15083\" src=\"https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-23.png\" srcset=\"https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-23.png 1558w, https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-23-300x153.png 300w, https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-23-1024x521.png 1024w, https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-23-768x390.png 768w, https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-23-1536x781.png 1536w\" sizes=\"auto, (max-width: 1558px) 100vw, 1558px\" \/><\/p>\n<p><a href=\"https:\/\/www.hitem3d.ai\/share\/bf95da8966bc4205a9af8343821514a9?inviteCode=z5Yks2FLJi&amp;shareType=history_model\"><strong>Examin 3D<\/strong><br \/>https:\/\/www.hitem3d.ai\/share\/bf95da8966bc4205a9af8343821514a9?inviteCode=z5Yks2FLJi&amp;shareType=history_model<\/a><\/p>\n<h1><strong>trellis 3D<\/strong><\/h1>\n<p>TRELLIS is a large 3D asset generation model. It takes in text or image prompts and generates high-quality 3D assets in various formats, such as Radiance Fields, 3D Gaussians, and meshes. The cornerstone of TRELLIS is a unified Structured LATent (SLAT) representation that allows decoding to different output formats and Rectified Flow Transformers tailored for SLAT as the powerful backbones. We provide large-scale pre-trained models with up to 2 billion parameters on a large 3D asset dataset of 500K diverse objects. TRELLIS significantly surpasses existing methods, including recent ones at similar scales, and showcases flexible output format selection and local 3D editing capabilities which were not offered by previous models.<\/p>\n<p>Source : https:\/\/github.com\/microsoft\/TRELLIS<\/p>\n<p><img loading=\"lazy\" decoding=\"async\" width=\"1477\" height=\"951\" class=\"wp-image-15084\" src=\"https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-24.png\" srcset=\"https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-24.png 1477w, https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-24-300x193.png 300w, https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-24-1024x659.png 1024w, https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-24-768x494.png 768w\" sizes=\"auto, (max-width: 1477px) 100vw, 1477px\" \/><\/p>\n<p><img loading=\"lazy\" decoding=\"async\" width=\"643\" height=\"355\" class=\"wp-image-15085\" src=\"https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-25.png\" srcset=\"https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-25.png 643w, https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-25-300x166.png 300w\" sizes=\"auto, (max-width: 643px) 100vw, 643px\" \/><\/p>\n<p><a href=\"https:\/\/fal.ai\/models\/fal-ai\/trellis?share=b60d48b7-6c61-4d17-84c8-bb5aa7877c08\"><strong>Examin 3D<\/strong><br \/>https:\/\/fal.ai\/models\/fal-ai\/trellis?share=b60d48b7-6c61-4d17-84c8-bb5aa7877c08<\/a><\/p>\n<h1><strong>Hunyuan 3D<\/strong><\/h1>\n<p><strong>HunyuanImage-3.0<\/strong>, which is a multimodal (2D) image generation model developed by Tencent . It is possible that this advanced model serves as the foundation for the 3D generation capabilities of the Hunyuan series.<\/p>\n<ul>\n<li><strong>Nature and Architecture:<\/strong> It is a native, open-source multimodal model that uses a unified autoregressive framework to directly integrate text and image, moving away from traditional DiT-based architectures .<\/li>\n<li><strong>Scale:<\/strong> It is the largest open-source Mixture of Experts (MoE) model for image generation, featuring 64 experts and a total of 80 billion parameters (with 13 billion activated per token) .<\/li>\n<li><strong>Features:<\/strong> It supports text-to-image and image-to-image generation workflows, including multi-image blending and multi-turn interactions .<\/li>\n<li><strong>Instruct Version:<\/strong> The \u00ab\u00a0HunyuanImage-3.0-Instruct\u00a0\u00bb version (introduced in January 2026) allows for enhanced creative image editing and complex prompt understanding, by breaking down queries and performing \u00ab\u00a0structured thinking\u00a0\u00bb to better grasp user intent .<\/li>\n<\/ul>\n<p>In Krea<\/p>\n<p><img loading=\"lazy\" decoding=\"async\" width=\"1735\" height=\"739\" class=\"wp-image-15086\" src=\"https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-26.png\" srcset=\"https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-26.png 1735w, https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-26-300x128.png 300w, https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-26-1024x436.png 1024w, https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-26-768x327.png 768w, https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-26-1536x654.png 1536w\" sizes=\"auto, (max-width: 1735px) 100vw, 1735px\" \/><\/p>\n<p>In <a href=\"https:\/\/hy-3d.net\/fr\">https:\/\/hy-3d.net\/fr<\/a><\/p>\n<p><img loading=\"lazy\" decoding=\"async\" width=\"1421\" height=\"1179\" class=\"wp-image-15087\" src=\"https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-27.png\" srcset=\"https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-27.png 1421w, https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-27-300x249.png 300w, https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-27-1024x850.png 1024w, https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-27-768x637.png 768w\" sizes=\"auto, (max-width: 1421px) 100vw, 1421px\" \/><\/p>\n<p><img loading=\"lazy\" decoding=\"async\" width=\"1004\" height=\"478\" class=\"wp-image-15088\" src=\"https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-28.png\" srcset=\"https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-28.png 1004w, https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-28-300x143.png 300w, https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-28-768x366.png 768w\" sizes=\"auto, (max-width: 1004px) 100vw, 1004px\" \/> <img loading=\"lazy\" decoding=\"async\" width=\"1032\" height=\"556\" class=\"wp-image-15089\" src=\"https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-29.png\" srcset=\"https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-29.png 1032w, https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-29-300x162.png 300w, https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-29-1024x552.png 1024w, https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-29-768x414.png 768w\" sizes=\"auto, (max-width: 1032px) 100vw, 1032px\" \/> <img loading=\"lazy\" decoding=\"async\" width=\"1103\" height=\"612\" class=\"wp-image-15090\" src=\"https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-30.png\" srcset=\"https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-30.png 1103w, https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-30-300x166.png 300w, https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-30-1024x568.png 1024w, https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-30-768x426.png 768w, https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-30-672x372.png 672w, https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-30-1038x576.png 1038w\" sizes=\"auto, (max-width: 1103px) 100vw, 1103px\" \/><\/p>\n<h1>Hyper3D<\/h1>\n<p>https:\/\/hyper3d.ai\/<\/p>\n<p><img loading=\"lazy\" decoding=\"async\" width=\"2048\" height=\"901\" class=\"wp-image-15091\" src=\"https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-31.png\" srcset=\"https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-31.png 2048w, https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-31-300x132.png 300w, https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-31-1024x451.png 1024w, https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-31-768x338.png 768w, https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-31-1536x676.png 1536w\" sizes=\"auto, (max-width: 2048px) 100vw, 2048px\" \/><\/p>\n<h1>Meta Sam3D<\/h1>\n<p><a href=\"https:\/\/aidemos.meta.com\/\">https:\/\/aidemos.meta.com\/<\/a><\/p>\n<p>SAM 3 is built as a unified, promptable model that enables segmentation with language, exemplars and visual prompts across images and videos. It leverages a large-scale, diverse training dataset and a powerful perception encoder backbone to achieve state-of-the-art performance in segmentation and tracking using open-vocabulary short text phrases and visual prompts.<\/p>\n<p><br \/>source : <a href=\"https:\/\/ai.meta.com\/research\/sam3\/\">https:\/\/ai.meta.com\/research\/sam3\/<\/a><\/p>\n<p><strong>Meta Sam3D<\/strong> is a spatial reconstruction tool related to Meta&rsquo;s <strong>Segment Anything Model 3 (SAM 3)<\/strong> .<\/p>\n<ul>\n<li><strong>Core Technology (SAM 3):<\/strong> SAM 3 is a unified AI model developed by Meta for precisely identifying, segmenting, and tracking objects in images and videos. It is a state-of-the-art model built on its predecessors (SAM 1 and SAM 2) .<\/li>\n<li><strong>User Interaction:<\/strong> Users can prompt SAM 3 using open-vocabulary text descriptions or visual examples to mask and track all matching object instances. It also features interactive capabilities, allowing developers to refine predictions with clicks and follow-up guidance .<\/li>\n<li><strong>SAM 3D:<\/strong> This is a related <strong>spatial-reconstruction tool<\/strong> that leverages the object segmentation and tracking power of SAM 3 to work in three dimensions .<\/li>\n<\/ul>\n<p><img loading=\"lazy\" decoding=\"async\" width=\"2048\" height=\"924\" class=\"wp-image-15092\" src=\"https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-32.png\" srcset=\"https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-32.png 2048w, https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-32-300x135.png 300w, https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-32-1024x462.png 1024w, https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-32-768x347.png 768w, https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-32-1536x693.png 1536w\" sizes=\"auto, (max-width: 2048px) 100vw, 2048px\" \/><\/p>\n<p><img loading=\"lazy\" decoding=\"async\" width=\"1024\" height=\"609\" class=\"wp-image-15093\" src=\"https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-33.png\" srcset=\"https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-33.png 1024w, https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-33-300x178.png 300w, https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-33-768x457.png 768w\" sizes=\"auto, (max-width: 1024px) 100vw, 1024px\" \/><\/p>\n<p><img loading=\"lazy\" decoding=\"async\" width=\"1122\" height=\"615\" class=\"wp-image-15094\" src=\"https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-34.png\" srcset=\"https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-34.png 1122w, https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-34-300x164.png 300w, https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-34-1024x561.png 1024w, https:\/\/www.keris-studio.fr\/blog\/wp-content\/word-image-15059-34-768x421.png 768w\" sizes=\"auto, (max-width: 1122px) 100vw, 1122px\" \/><\/p>\n\n\n<figure class=\"wp-block-embed is-type-video is-provider-youtube wp-block-embed-youtube wp-embed-aspect-16-9 wp-has-aspect-ratio\"><div class=\"wp-block-embed__wrapper\">\n<iframe loading=\"lazy\" title=\"Slowly turn around the building 202605141508\" width=\"474\" height=\"267\" src=\"https:\/\/www.youtube.com\/embed\/__ye8DVW1rg?feature=oembed\" frameborder=\"0\" allow=\"accelerometer; autoplay; clipboard-write; encrypted-media; gyroscope; picture-in-picture; web-share\" referrerpolicy=\"strict-origin-when-cross-origin\" allowfullscreen><\/iframe>\n<\/div><\/figure>\n\n\n\n<figure class=\"wp-block-embed is-type-video is-provider-youtube wp-block-embed-youtube wp-embed-aspect-16-9 wp-has-aspect-ratio\"><div class=\"wp-block-embed__wrapper\">\n<iframe loading=\"lazy\" title=\"Fly down to building 202605161303\" width=\"474\" height=\"267\" src=\"https:\/\/www.youtube.com\/embed\/svjuFbmXkJ8?feature=oembed\" frameborder=\"0\" allow=\"accelerometer; autoplay; clipboard-write; encrypted-media; gyroscope; picture-in-picture; web-share\" referrerpolicy=\"strict-origin-when-cross-origin\" allowfullscreen><\/iframe>\n<\/div><\/figure>\n\n\n","protected":false},"excerpt":{"rendered":"<p>Starting base This image serves as the foundation from which the AI extrapolates the three-dimensional geometry (mesh), vertices, and topology of the final 3D model.<\/p>\n","protected":false},"author":2,"featured_media":15098,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[6,190,593,8,26],"tags":[625,636,543,386],"class_list":["post-15059","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-3d","category-architecture-2","category-artificial","category-methodologie","category-production","tag-ia-generative","tag-image2-3d","tag-intelligence-artificielle","tag-methodology"],"_links":{"self":[{"href":"https:\/\/www.keris-studio.fr\/blog\/index.php?rest_route=\/wp\/v2\/posts\/15059","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.keris-studio.fr\/blog\/index.php?rest_route=\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.keris-studio.fr\/blog\/index.php?rest_route=\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.keris-studio.fr\/blog\/index.php?rest_route=\/wp\/v2\/users\/2"}],"replies":[{"embeddable":true,"href":"https:\/\/www.keris-studio.fr\/blog\/index.php?rest_route=%2Fwp%2Fv2%2Fcomments&post=15059"}],"version-history":[{"count":4,"href":"https:\/\/www.keris-studio.fr\/blog\/index.php?rest_route=\/wp\/v2\/posts\/15059\/revisions"}],"predecessor-version":[{"id":15103,"href":"https:\/\/www.keris-studio.fr\/blog\/index.php?rest_route=\/wp\/v2\/posts\/15059\/revisions\/15103"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.keris-studio.fr\/blog\/index.php?rest_route=\/wp\/v2\/media\/15098"}],"wp:attachment":[{"href":"https:\/\/www.keris-studio.fr\/blog\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=15059"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.keris-studio.fr\/blog\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=15059"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.keris-studio.fr\/blog\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=15059"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}