Glm-5.3-flash max on ZAI coding plan + opencode
It's pretty cool for something that runs on DGX Spark x2. Considering the speed, the local candidate is Qwen-3.8-Next..?
Running brand-new Qwen 3.8 Flash Next locally on a MacBook 64GB M5 Max at a smooth 30 tok/s.
Full agent workflow web search, Python, multi-table research report completed in under 8 minutes.
And on a real phone? No dgx no GPU only CPU?
Imagine running Qwen 3.8 Flash Next on a 12GB Android phone
And phone. Imagine running Qwen 3.8 Flash Next on a 12GB Android phone - 2 tok/s
You have reached the end of the archive
All of qwen38