Mistral Medium 3.5
Mistral Medium 3.5 is a 128B-parameter model you reach through an API — too large to self-host, so this page covers what it does, how to access it, and what to run locally instead.
You don't run Mistral Medium 3.5 locally
At 128B parameters, a Q4-class build of Mistral Medium 3.5 would need roughly 77 GB — a 256GB Mac Studio could hold the weights in principle, but no registry-verified local build exists today (0.6 GB per billion parameters, our standard Q4 rule). Access is through the hosted API.
Apr 30, 2026. 128B dense. SWE-Bench Verified 77.6%. Merges Magistral + Devstral 2 + Pixtral. GGUF available but 96GB+ Mac required. Open weights: mistralai/Mistral-Medium-3.5-128B on Hugging Face (modified MIT).
Strong local alternatives
More Mistral models
Frequently asked questions
Can I run Mistral Medium 3.5 locally?
Not realistically. Mistral Medium 3.5 is a 128B-parameter model; a Q4-class build would need roughly 77 GB — a 256GB Mac Studio could hold the weights in principle, but no registry-verified local build exists today. The hosted API or a smaller open model is the practical path.
How do I access Mistral Medium 3.5?
Through the vendor-hosted API. See the official source linked on this page.
What is the best local alternative to Mistral Medium 3.5?
Qwen3 235B A22B is the strongest local model we track (235B). It runs on a single high-memory Mac or GPU; see its page for exact hardware.
Cite this page
ModelFit: Mistral Medium 3.5 — specs, memory math and hardware verdicts. https://modelfit.io/models/mistral-medium-3.5/ (dataset updated 2026-09-03, CC BY 4.0).