Unverified50% confidenceFactExact time
The developer fine-tuned general coding capability based on Qwen3.5-4B, with Three.js scene generation as the core test task.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
10/27/2026
First Seen
Valid until: 1/25/2027
Sources
4B Model Fine-Tuning in Practice: Three.js 3D Scene Generation Keeps Improving
redditr/LocalLLaMA10/26/2026
Related Entities
Related Claims
UnverifiedA 4B-level model reaching a usable level in Three.js structured code generation means independent developers and small teams can build 3D scene prototyping tools at extremely low cost.67% similarUnverifiedJS 3D Builder (Three.js) is a benchmark whose task is, starting from written architectural specifications, to procedurally generate 3D building code using Three.js, scored by server-side static analysis, without executing the code or using an LLM judge.65% similarUnverifiedWhen generating structured support tickets in the browser with Qwen2.5-0.5B-Instruct, the biggest difficulty is getting the model to reliably output correctly formatted JSON, rather than understanding the complaint content.64% similarUnverifiedThe large-model test of "writing code to generate 3D content" examines the model's ability to generate runnable HTML/JavaScript code (often with the help of WebGL libraries such as Three.js) and have 3D scenes rendered in real time by the browser.62% similarUnverified7B-class code models (such as Qwen2.5-Coder 7B and DeepSeek-Coder) can understand common HTML/CSS/JavaScript requirements and generate basic component code61% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1102266API
curl https://kongchang.com/api/v1/knowledge/claims/1102266MCP
get_claim(id=1102266)