inference speed matters more than benchmark scores for local models
r/LocalLLaMA
•
AI Research
After testing a bunch of local models for actual coding tasks i've come to the