Other models of this size have done well in the past. Qwen2.5 coder, for example, can correctly answer the question at 7B. Deepseek R1 was also capable of giv

Other models of this size have done well in the past. Qwen2.5 coder, for example, can correctly answer the question at 7B. Deepseek R1 was also capable of giving a correct response. It's obviously a doable. Such a model locally is useful in autocomplete while programming.