Back to AIBriefs
AnalysisAI Models

Local LLMs now reach ~75% of frontier model speed for coding

Local LLMs now reach ~75% of frontier model speed for codingVicki Boykis

Vicki Boykis reports that local models, especially Gemma 4, now achieve about 75% the accuracy and speed of frontier models for agentic coding tasks. She uses a 2022 M2 Mac with 64GB RAM and tools like LM Studio and llama.cpp.

Jun 15, 1:00 PM
Local LLMs now reach ~75% of frontier model speed for coding — AIBriefs