Summary
According to a post on X by @art3dlabs, which reposts a note from dextune, CCORE's Qwen3.5-4B-GLB-3D-Coder moved past preview on Nov 5. The release includes Q4_K_M, Q8_0 and Safetensors weights for the 4B model, which is described as writing Three.js pages with 3D previews and GLB export for editors such as Godot. The model is licensed under Apache 2.0. The details come from the release note as relayed in the post, and the poster says they did not run the weights themselves.
AI-assisted summary based on the listed source.
What happened
List member art3dlabs reposted dextune's note that CCORE's Qwen3.5-4B-GLB-3D-Coder moved past preview on Nov 5 with Q4_K_M, Q8_0, and Safetensors weights — a 4B model that writes Three.js pages with 3D previews and GLB export for editors like Godot, Apache 2.0 (original about 162 likes, 19 reposts, 146 bookmarks...
Why it matters
The model targets developers and designers who build 3D content for the web or game engines, and the published weights let them test it under a permissive license. Its usefulness will depend on how it performs on real 3D tasks, which this post does not show. Watch for independent tests and community reports on output quality from the released weights.
Signal Intelligence
Signal Strength 90%
Technical label WATCH
Public Interest 33
Category PRODUCT UPDATE
Reader Depth TECHNICAL
Signal Strength reflects source quality, relevance, freshness and evidence. Public Interest helps organize discovery; it is not proof of truth.
Public Interest components
Recognizable Entity Score 0
Practical Impact Score 46
Novelty Interest Score 72
Consequence Score 18
Curiosity Score 16
Shareability Score 49
Why this is here
VQV surfaced this signal because it is recent, relevant to Open Source LLMs, connected to Scobleizer News Pulse (X).
