Hyper3D's Rodin Gen-2.5 AI now handles mesh repair, part splitting, and multi-color assignment automatically before a file ever reaches the slicer.
AI-generated models still need prep work
Generating a 3D model with AI takes seconds. Preparing that model for actual printing is a different story. Meshes can have broken topology, non-manifold edges, or too many faces for a slicer to handle. Complex objects often need to be split into parts that fit on the build plate. Multi-color prints require separate mesh regions for each material.
Hyper3D, the generative 3D platform built by Deemos, used Formnext Asia Shenzhen to show that AI can take on those steps too. The company demonstrated new print preparation tools built around its Rodin Gen-2.5 model, which launched back in May 2026.
Rodin Gen-2.5 generates up to 10 million faces
Rodin Gen-2.5 offers five geometry quality tiers. Extreme-Low mode produces lightweight models in roughly four seconds for fast ideation. Extreme-High mode outputs meshes with up to 10 million faces, capturing fine surface details that actually reproduce in a physical print.
More faces is not always better. A 10 million face mesh carries a heavy processing cost. For most consumer prints, the Medium tier hits the right balance between detail and slicer performance. The higher tiers suit sculptures, collectibles, and high-resolution resin work.
Qixuan Zhang, CTO of Deemos, described Rodin Gen-2.5 as moving from million-polygon generation in seconds to raw outputs exceeding 10 million polygons. That range gives users a wider span of geometric fidelity depending on what the final print demands.
Print preparation tools are the real update
The headline feature at Formnext Asia was not the model generator itself. It was the preparation pipeline Hyper3D built on top of it. The system can automatically repair messy or broken meshes, edit specific regions of a finished model using a text prompt, and split complex objects into separate printable parts that fit within the build volume.
For multi-color workflows, the software assigns mesh regions to specific colors before export, cutting out a manual step that most slicers still require users to do by hand.
Why this matters for 3D printing workflows
The bottleneck in AI-to-print pipelines has never been model generation. It is what comes after. A model that looks clean in a viewport can still fail during slicing or printing if the mesh structure is wrong. Hyper3D's new tools target that gap directly.
Rodin Gen-2.5 is available now through the Hyper3D website, API, and enterprise deployment options. The print preparation features were shown publicly at Formnext Asia Shenzhen 2026 and are expected to roll out to workspace users in the coming months.
Comments (0)
No comments yet. Be the first!
Leave a Comment