| ▲ | kadoban 5 hours ago | |
Why are the models being used so old (in llm years anyway). Why no Qwen 3.8 ? | ||
| ▲ | droidjj 5 hours ago | parent | next [-] | |
Based on how vibe-coded this whole thing is, I’m guessing the newer models weren’t in the training data and the agent just picked what it knew existed. | ||
| ▲ | x-complexity 4 hours ago | parent | prev [-] | |
Qwen 3.6 & 3.8 have the exact same model architecture. In theory, it should just be a drop-in replacement. | ||