😃Made My Own Invoicing Dekstop App With Qwen 3.8 27B+OpenCode+LM Studio On RTX 5090..Saved Me $100 a Year I Was Paying

My first fully vibe coded Desktop app with Qwen 3.8 27B as someone with near 0 coding knowledge lol. Was paying $100 a year for an online Invoicing App so i decided to cook up one for myself. The results are mind blowing..I basically one shotted everything throughout development & implementation as Open Code's coding agents did all the work, fixing bugs, web search etc. All i did is connect my Qwen Model to Open Code & started coding away. I was mostly screenshoting and telling the model what i wanted, it even created in app tools to import all my 5 year+ documents & info which all worked. The New Open Code v2.0 is crazy good, couldn't do any of this months ago without manually download a crap tone of plug-ins and tools. My context window always stayed between 57-90K thanks to Open Code's auto Compaction..very happy with the results. PC Setup AMD Ryzen 9 9950X3D RTX 5090 32GB 96GB RAM DDR5 Workflow Setup(31GB Vram Usage) AI Model - Qwen 3.8 27B Q4_K_M GGUF(120K Context) Harness - Open Code 2 (Default Settings) Model Manger - LM Studio(Default Model Settings, Thinking On) Model Performance/Project Token Usage Average Token/sec - 80-90 Input Tokens - 294 Million Output Tokens - 1.26 Million I Quick used a bit of Qwen 3.6 35B A3B..Q5 K M for the last feature implementation just to test, gave me 220token/s and worked well too, the rest of Dev was on Qwen 3.8 27b. All this was done on my RTX 5090, Model and 120K Context all fit perfectly in Vram. Can't wait for Qwen 4 and it's improvements as well as Open Code's future updates, will try to experiment with other Tools & Skills cause i only used Open Code's default setup.