Leaderboard/Qwen3-32B-4bit/M3 Ultra (60-core GPU)/r_anmmc80-aoq

Qwen3-32B-4bit on M3 Ultra (60-core GPU) + 96GB unified

Name: Qwen3-32B-4bit on M3 Ultra (60-core GPU) + 96GB unified
Creator: llm-speed
License: https://www.apache.org/licenses/LICENSE-2.0
Keywords: LLM benchmark, tokens per second, Qwen3-32B-4bit, M3 Ultra (60-core GPU) + 96GB unified, mlx

M3 Ultra (60-core GPU) + 96GB unified

suite suite-v1

cli 0.0.1-dev

signedG8xb3zMu3+…

Embed badgesubmitted Apr 28, 2026

https://llm-speed.com/r/r_anmmc80-aoq

Workload results

Workload	Backend	Model	decode tok/s	prefill tok/s	TTFT	p50	p95
chat-short	mlx@0.31.3	mlx-community/Qwen3-32B-4bit	34.41tok/s	95.12tok/s	1,156ms	29.1ms	29.2ms

Reproduce on your machine

Same workload, same model, signed at your rig. The exact command that produced this run:

$ pipx install llm-speed && llm-speed bench --model 'mlx-community/Qwen3-32B-4bit' --workload 'chat-short'

Runs in about a minute. Your number lands on the leaderboard signed and linkable. How it's measured.

Embed this run

Drop the badge into a README, blog post, or signature. Each render is a backlink to the signed result.

[![llm-speed: 34.4 tok/s on M3 Ultra (60-core GPU) (Qwen3-32B-4bit)](https://llm-speed.com/badge/r_anmmc80-aoq.svg)](https://llm-speed.com/r/r_anmmc80-aoq)

Related benchmarks

More Qwen3-32B-4bit benchmarks — every backend and rig that has run this model.
More M3 Ultra (60-core GPU) LLM benchmarks — every model measured on this hardware.

Provenance

Run ID: r_anmmc80-aoq
Fingerprint hash: bbf15132ccbbe7d7
Public key: G8xb3zMu3+pznEici/TiW0gPk5qSNIYIikGCwm1rMdQ=
Received: 2026-04-28 14:19:50