Question: Commercial use of LingBot-VLA 1.0 — dependency on Qwen2.5-VL-3B-Instruct license

[nguyetnga08bb88-cpu]

Hi LingBot-VLA team,
First of all, thank you for open-sourcing this great work. I have a question regarding commercial use that I hope you can clarify.

Background

• The LingBot-VLA 1.0 codebase and weights are released under Apache-2.0, which is commercial-friendly.
• However, according to the README, training and inference require the Qwen2.5-VL-3B-Instruct checkpoint as the VLM backbone.
• Based on public reports (e.g., The Batch / deeplearning.ai) and the Qwen Research License, Qwen2.5-VL-3B appears to be restricted to non-commercial use only, unlike the 7B/32B variants which are Apache-2.0.

Questions

1. If a company wants to use LingBot-VLA 1.0 (4B) in a commercial product, does the dependency on Qwen2.5-VL-3B-Instruct impose a non-commercial restriction?
2. If yes, is there a recommended path — e.g., obtaining a commercial license from Alibaba for Qwen2.5-VL-3B, or swapping the backbone to a commercially permissive VLM?
3. Does LingBot-VLA 2.0 (6B) avoid this issue entirely, since it uses Qwen3-VL-4B-Instruct (which appears to be Apache-2.0)?

I want to make sure I understand the licensing correctly before building a commercial application on top of this project. Any clarification would be greatly appreciated.
Thanks!