Qwen3.5-4B-GLB-3D-Coder builds a 3D tank in 250 s on my 16 GB RTX 5060 Ti
This is a dev post classified by AGK Intelligence as AI dev tools (a tool drop), kept by the Dev Radar because it carries real work, not commentary.
This could be so cool. Qwen3.5-4B-GLB-3D-Coder builds a 3D tank in 250 s on my 16 GB RTX 5060 Ti Median of 6 models, one prompt each: 247 s per model 18,678 tokens per model 75.7 tok/s 8.58 GiB VRAM 4 of 6 valid on the first try Q8_0 (8.5 bpw) from ccore, llama.cpp b12124 131072 context, thinking on, no token limit One try per prompt, no help, no fixes Temperature 0.8, top-p 0.95, seed 42, as on the model card The motorcycle and the fighter jet failed. The page code the model wrote had a bug (a variable declared twice, a function never defined), so no 3D file came out. I counted both as f
Posted by Niklas (2.8k followers) 2 h ago · 4 likes · 208 views · view the original post on X. Kept by the Dev Radar as AI dev tools. Tools mentioned: qwen3.5-4b-glb-3d-coder.
More dev work like this
- Soniox STT v6 delivers our most accurate and fastest voice typing yet across 60+… — @soniox_ai
- NXP i.MX95 export for Ultralytics YOLO has shipped! 🚀 — @ultralytics
- Your agent wrote the report. Now give it a layout. — @DanKornas
- Introducing our next generation of model routers: — @OpenRouter
- "Sovereignty means ownership over the value that your organization creates.” — @PalantirTech
- The first AI agent companies to try Daytona said the product was useless. — @guilleflorvs
- A model that spends tokens reasoning can help train another model to make decisions… — @techNmak
- thinking machines dropped new model — @notjazii
Every post is read and classified by AGK Intelligence: what it is, which market it belongs to, and whether the link is a real tool. 31.8k posts from 5k X accounts over the last 21 days, 9.5k tools, 12 markets. New posts are collected several times an hour and the full ranking is recomputed every hour — last update 2026-11-06 16:54 UTC. Full method.