Fast local LLMs on Apple Silicon: sub-second model loads, faster than Ollama on long prompts. OpenAI and Ollama-compatible.https://squish.run/ https://github.com/konjoai/squishen Sun, 28 Jun 2026 11:56:43 -0000 Sun, 28 Jun 2026 11:56:43 -0000 1440 MkDocs RSS plugin - v1.19.0 None https://squish.run/ Wesley Scholl Benchmarks apple-silicon benchmarks local-llm mlx quantization rust A local LLM inference server for Apple Silicon, up to 9.8× faster than Ollama on long prompts, with the honest benchmarks. https://squish.run/blog/local-llm-fast-enough/ Fri, 26 Jun 2026 03:16:00 +0000 Squishhttps://squish.run/blog/local-llm-fast-enough/