Those of you who pay attention to AI developers on X and other platforms know they are always looking for ways to run powerful models locally. It is possible to run Codex on your Mac with a local model like Qwen3.6‑35B‑A3B 6‑bit on an M5 Max 128GB machine. To do something like this, you will need to point Codex’s backend to LM Studio’s OpenAI compatible server and tell Codex to use that provider and model.
Video on LM Studio MLX v.1.8.5 + Codex + qwen3.6-35b-a3b 6bit on an M5 Max 128GB.
Video is in normal speed to give you an idea of the experience.Command used:
codex –oss -m qwen/qwen3.6-35b-a3bLocal AI is becoming powerful 💪 pic.twitter.com/nECLMRtgW8
— Ivan Fioravanti ᯅ (@ivanfioravanti) June 6, 2026
You can use a command like this to tell Codex to use LM Studio as a provider.
codex --oss -m qwen/qwen3.6-35b-a3b
The days of having to set up boundary wires to keep your robot mowers in…
Many of us catch up to our favorite podcasts and audio programs in bed. Wearing…
There is a good reason many companies use E-ink in their devices. This technology is…
In the past few months, we have covered a bunch of SSDs that combine with…
Here is another 6K camera designed with serious content creators in mind. The Kinefinity Vista…
When it comes to productivity, multiple screens are better than one. Take the NexFold Fold7…