---
title: "Ollama is Too Slow: Try This Instead!"
source_url: https://youtu.be/9LTe1RC3hj0
captured_at: 2026-09-06
author: "Eric Tech"
contributor: "ชิดชนก"
platform: youtube
---
# Ollama is Too Slow: Try This Instead!

youtube video by **Eric Tech**, 10m44s, uploaded 2026-05-26. URL: https://youtu.be/9LTe1RC3hj0
Tags: claude code, oMLX, ollama, claude code local, claude code mac, run claude code locally, local llm mac, local ai mac, claude code free, mlx, local llm, qwen 3.5, gemma 4, claude code setup, claude code tutorial

## Shared in LINE Note+
- 2026-09-06 05:25 by ชิดชนก

## Description
Try SureThing for Free👉 

Tired of Claude Code feeling slow when you run it locally on Mac? In this video I show you how to use oMLX to run local AI models 4x faster than Ollama on Apple Silicon, then plug them straight into Claude Code.

Key takeaways:

- Install oMLX in 2 minutes and run Qwen 3.5 9B or Gemma 4 locally on Mac without crashes
- Use the --bare flag to fit Claude Code's heavy system prompt inside small local models
- Side-by-side benchmark: oMLX vs Ollama on the same MacBook with the same model.

🔗 Join our School community: skool.com/erictech

🔗 Check out bookzero.ai — AI-powered bookkeeping built entirely with Claude Code

Checkout Testsprite:  

📌 Mentioned playlist:

- Claude Code For Free, Every Single Method: 


Timestamps:
0:00 Intro
1:18 Install oMLX
2:23 Server Setup
2:50 Download Models
3:44 Test Chat (Qwen 3.5)
4:23 Ollama Fails (Qwen 3.5)
5:30 Launch Claude Code
5:55 Bare Flag
6:42 Speed Test
7:31 Sponsor
9:25 Final Results
10:01 Outro

#claudecode #aitools #softwaredevelopment