Comment by Aurornis

In the local LLM communities there is a lot of respect for the Qwen models, but everyone comes to acknowledge that they do a lot of benchmaxxing after using them. Even at full precision they're never as good as models with similar benchmarks.