Joined November 2022
- Tweets 1,618
- Following 800
- Followers 15,113
- Likes 3,924
513 Photos and videos
Pinned Tweet
May 19
Introducing #Rodin Gen-2.5🚀
🔥World’s 1st 10 MILLION polygon #3D GenAI — down to skin microstructures.
1️⃣1M-poly in 4s
2️⃣Adaptive thinking effort - just like LLMs
3️⃣3D-native textures, no blind spot
4️⃣Batch up to 10 results
💥Manual BANG to Parts & more...
🚨$1 for first mo!
44
128
925
571,428
Hyper3D by Deemos retweeted
Jun 8
Indie Game Developer @alexfredo87 combined Rodin Gen-2.5 Extreme-High tier with a custom Skin Shader, and it achieves really impressive results.
May 19
Introducing #Rodin Gen-2.5🚀
🔥World’s 1st 10 MILLION polygon #3D GenAI — down to skin microstructures.
1️⃣1M-poly in 4s
2️⃣Adaptive thinking effort - just like LLMs
3️⃣3D-native textures, no blind spot
4️⃣Batch up to 10 results
💥Manual BANG to Parts & more...
🚨$1 for first mo!
1
14
112
13,018
Hyper3D by Deemos retweeted
Jun 10
Claude Fable made this game with Rodin Gen-2.5 in a one-line prompt.
The most exciting part is, it did not use BANG to Parts skill; it split the model to make animations by itself🤯
May 19
Introducing #Rodin Gen-2.5🚀
🔥World’s 1st 10 MILLION polygon #3D GenAI — down to skin microstructures.
1️⃣1M-poly in 4s
2️⃣Adaptive thinking effort - just like LLMs
3️⃣3D-native textures, no blind spot
4️⃣Batch up to 10 results
💥Manual BANG to Parts & more...
🚨$1 for first mo!
2
10
66
9,550
Hyper3D by Deemos retweeted
Jun 10
Your first 3D generation in Leonardo, made simple.
Four settings. One clear starting point. Everything else is yours to explore.
Try 3D in Leonardo today.
8
5
25
2,094
Hyper3D by Deemos retweeted
Jun 6
Paper Alert🚨#CVPR 2026
🤔 How does a #3D model understand that a guitar should naturally lean against a bookcase like this?
🚀Introducing PARSE: a framework for procedurally generating spatially grounded 3D scenes from part-level relational descriptions.
More details👇
7
12
116
12,780
Hyper3D by Deemos retweeted
Jun 6
I’m testing Rodin Gen-2.5 Extreme-High and its 12K texture with a custom skin shader I created. The result is pretty good. @DeemosTech
3
6
106
7,521
Jun 6
(3/4) To support spatial reasoning, we built PARSE-10K—a dataset of 10,000 indoor scenes with detailed part-level annotations. It helps VLMs better understand spatial relationships and generate more realistic 3D scenes.
1
6
741
Jun 6
(4/4)
🔍Paper: arxiv.org/abs/2603.07704
🌍Project Page: otanaaa.github.io/PARSE-proj…
🚀Coming Soon to hyper3d.ai
9
6,818
Rodin Gen-2.5 by @DeemosTech is now live in Comfy!
Production-scale 3D generation with controllable speed, detail, and texture quality.
→ Generate 3D assets from text prompts or reference images
→ Choose generation tiers for different workflows, from fast drafts to extreme-high fidelity
→ Control mesh density, mesh type, texture quality, and material output
→ Create 3D-native textures with full-surface coverage and no blind spots
Try the link below to the Workflow 👇
11
25
277
17,673
Jun 5
Testing Rodin Gen-2.5 Extreme-High
It's so crazy!!!!
May 19
Introducing #Rodin Gen-2.5🚀
🔥World’s 1st 10 MILLION polygon #3D GenAI — down to skin microstructures.
1️⃣1M-poly in 4s
2️⃣Adaptive thinking effort - just like LLMs
3️⃣3D-native textures, no blind spot
4️⃣Batch up to 10 results
💥Manual BANG to Parts & more...
🚨$1 for first mo!
7
18
214
15,076
Jun 4
Come check out Kinematify at #ICRA2026 today 👀
📍 Hall C, Poster #333
From RGB images or text to high-DoF articulated objects with physically consistent constraints, Kinematify brings 3D generation one step closer to robotics-ready assets.
Come stop by and chat with us!
10 Nov 2025
Paper Alert🚨
🤔Generate interactive Asset after BANG to Parts?
Introducing our recent work, Kinematify: Open-Vocabulary Synthesis of High-DoF Articulated Objects—export URDF for motion planning, sims & training.
🎮Demo: sites.google.com/deemos.com/…
📖Paper: arxiv.org/abs/2511.01294
4
3
25
3,788
Hyper3D by Deemos retweeted
Jun 4
Rodin V2 on LeonardoAI is seriously strong for 3D assets.
One reference image → usable 3D → refine and export ✅
Jun 3
One image is enough
You can now turn a single character or object into clean reference angles for a full 3D workflow, without building a long prompt or character sheet
Generate the views, move into Rodin V2 on LeonardoAI, and give it real 3D depth from there
Let me show you how it works:
5
7
65
6,142
Jun 4
3D Head Modeling Test
Input Image from @virtualfilmer
May 19
Introducing #Rodin Gen-2.5🚀
🔥World’s 1st 10 MILLION polygon #3D GenAI — down to skin microstructures.
1️⃣1M-poly in 4s
2️⃣Adaptive thinking effort - just like LLMs
3️⃣3D-native textures, no blind spot
4️⃣Batch up to 10 results
💥Manual BANG to Parts & more...
🚨$1 for first mo!
13
42
519
50,661
Jun 3
🤯Single Image to a Complex #3D Scene.
3D Native Texture is one of the biggest updates in #Rodin Gen-2.5.
No more texture blind spots from complex geometry.
Preserve the original shading directly from the input image — even Toon Shading.
May 19
Introducing #Rodin Gen-2.5🚀
🔥World’s 1st 10 MILLION polygon #3D GenAI — down to skin microstructures.
1️⃣1M-poly in 4s
2️⃣Adaptive thinking effort - just like LLMs
3️⃣3D-native textures, no blind spot
4️⃣Batch up to 10 results
💥Manual BANG to Parts & more...
🚨$1 for first mo!
2
4
57
5,582
We interviewed the CTO of @DeemosTech #Hyper3D to learn all about the latest Gen-2.5 version of #Rodin, which includes controllable #3D #AI workflows and Smart Low-Poly topology generation.
Full interview: 80.lv/articles/how-hyper3d-r…
4
10
3,590