Run Qwen3.8-27B on M3 Ultra

Qwen3.8-27B runs at 38 tok/s (37.6 as recorded) on a 512 GB M3 Ultra, a community run reported by github.com (MLX, 4-bit, 512-token prompt, 128 generated). It needs about 19 GB of unified memory. It is Alibaba's 27.8B dense model, released under the Apache 2.0 licence.

Speed

community38

tok/s

First Token

—

seconds

RAM Needed

GB minimum

Engine

MLX

of this figure

Want to run Qwen3.8-27B faster — or step up to bigger models? Find your Mac →

The figure

This is a community run: a real run, reported by its source and linked below. The index records it as given and never adjusts it.

Metric	Value
Tokens per second	community38 tok/s
Provenance	community run · github.com
Quantization	4-bit
Engine	MLX
Context	512-token prompt, 128 generated
Figure dated	2026-09-20
Minimum RAM	19 GB
Parameters	27.8B
Chip bandwidth	819 GB/s

Every figure for Qwen3.8-27B on the M3 Ultra

Each row as the index holds it, with its source. “Mac RAM” is the memory of the Mac a run used, not the model’s requirement (19 GB). The headline figure follows the reference rule: the lower-median measured plain-decode run at 5 bits or fewer (a run whose quant was not stated counts only when no stated-quant run qualifies), else the estimate; a run marked “not used to anchor estimates” is listed here but never becomes the headline.

Speed	Quant	Engine	Mac RAM	Context	Source	Note
sourced40 tok/s	Q4_K_M	LM Studio	96 GB	—	bgr.com	—
community37.6 tok/s	4-bit	MLX	512 GB	512-token prompt, 128 generated	github.com	Mac Studio, 80-core GPU, 512 GB, macOS 26.5.1. mlx_lm.benchmark -p 512 -g 128 -n 5 (mlx-lm 0.31.3, MLX 0.32.2): plain decode with no HTTP server or chat template, median of the five trial lines in the repo's engine.json; mlx-community/Qwen3.8-27B-4bit. The repo's headline test streams the same cell through mlx_lm.server (thinking and prompt cache off): 32.8 tok/s. The source discloses a stuck background process on one of the M3 Ultra's 32 CPU cores during these runs, which can only have slowed it.
sourced31 tok/s	oQ4e	MLX	512 GB	6,091-token prompt	macstories.net	MacStories review: oMLX 0.7.0.dev2, MTP off, thinking off, single run; Qwen3.8-27B-oQ4e-mtp checkpoint.
sourced12.46 tok/s	BF16	MLX	512 GB	8,192-token context, 128 generated	machinelearning.apple.com	Apple ML Research: mlx_lm.benchmark -p 8192 -g 128 -b 1 on one Mac Studio.

Provenance key

sourced

Published figure — vendor or press review

estimated

Derived from published bandwidth model

community

Linked individual run — blog, forum or submission

Setup: run Qwen3.8-27B on a Mac

Ollama — install it (brew install ollama or the app from ollama.com), then pull the model by its tag, checked on ollama.com on 2026-10-24:

ollama run qwen3.8:27b

MLX — pick a build from Hugging Face’s MLX listings, then run it with its repo id:

pip install -U mlx-lm

mlx_lm.generate --model <repo-id> --prompt "Hello"

Downloads: Search the Ollama library → · model card →

The 38 tok/s figure above came from MLX at 4-bit; other runtimes and quants run at different speeds.

Prefer a GUI? LM Studio downloads GGUF and MLX builds in one click. Step by step: install Ollama on a Mac.

Qwen3.8-27B on other Apple Silicon chips

Chip	Speed	Quant	Engine	Bandwidth
M5 Ultra	sourced48 tok/s	oQ4e	MLX	1228 GB/s
M5 Max	estimated29 tok/s	Q4_K_M	MLX	614 GB/s
M4 Max	estimated26 tok/s	Q4_K_M	MLX	546 GB/s
M3 Max	estimated19 tok/s	Q4_K_M	MLX	400 GB/s
M5 Pro	sourced17 tok/s	4-bit	llama.cpp	307 GB/s
M4 Pro	estimated13 tok/s	Q4_K_M	Ollama	273 GB/s
M6	community9 tok/s	QAT	LM Studio	170 GB/s

About Qwen3.8-27B

Alibaba's Apache 2.0 27.8B dense vision-language model (2026-08-14): image and video input, 262K context (1M with YaRN), thinking mode with reasoning_effort. Runs 4-bit in ~19 GB, so it fits a 24 GB Mac; ~29 tok/s estimated on M5 Max, and BGR's review measured ~55 on an M5 Ultra and ~40 on an M3 Ultra — within 4% of the index's estimates. Model card: SWE-Bench Pro 61.7 (vendor scaffold), Terminal-Bench 2.1 73.0, LiveCodeBench v6 90.3, GPQA 89.2. 1,000+ GGUF quants; Ollama, llama.cpp and LM Studio supported.

Apache 2.0 licence · 27.8B · 262K context · LLMCheck capability 46/50 · see its leaderboard rank →

System requirements

• 19 GB of unified memory or more (the index’s minimum: the weights at the quant that fits, plus ~2 GB)
• About 17 GB of free disk space for the model file
• Any Apple Silicon Mac with that much memory runs it; the M3 Ultra’s 819 GB/s memory bandwidth sets the speed shown here
• A current runtime: Ollama, LM Studio, llama.cpp or MLX (Ollama install guide)

Macs with the M3 Ultra

Mac Studio M3 Ultra (256 GB) fits · 819 GB/s
Mac Studio M3 Ultra (512 GB) fits · 819 GB/s
Best local LLMs for a 24 GB Mac the smallest RAM tier that fits Qwen3.8-27B

🛒 Get a Mac that runs Qwen3.8-27B

Qwen3.8-27B needs about 19 GB of unified memory. These current Apple Silicon Macs have the headroom to run it comfortably:

Mac mini M6 (32GB) on Amazon → Mac mini M5 Pro on Amazon →

As an Amazon Associate, LLMCheck earns from qualifying purchases. These affiliate links cost you nothing extra and help keep the LLMCheck index free.

Compare More Models

See how Qwen3.8-27B stacks up against other models on your specific Mac hardware.

Open Compare Tool Full Leaderboard