Run Qwen 3 8B on M5 Ultra

Qwen 3 8B runs at 149 tok/s on a 256 GB M5 Ultra, a published figure reported by vsesmart.cz (llama.cpp, 4-bit, empty context, 128 generated). It needs about 6 GB of unified memory. It is Alibaba's 8B dense model, released under the Apache 2.0 licence.

Speed
sourced149
tok/s
First Token
—
seconds
RAM Needed
6
GB minimum
Engine
llama.cpp
of this figure
Want to run Qwen 3 8B faster — or step up to bigger models? Find your Mac →

The figure

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

MetricValue
Tokens per secondsourced149 tok/s
Provenancepublished figure · vsesmart.cz
Quantization4-bit
Enginellama.cpp
Prompt processing (prefill)sourced4205 tok/s at 4K context
Contextempty context, 128 generated
Figure dated2026-10-26
Minimum RAM6 GB
Parameters8B
Chip bandwidth1228 GB/s

Every figure for Qwen 3 8B on the M5 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 (6 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.

SpeedQuantEngineMac RAMContextSourceNote
sourced176 tok/s4-bitMLX256 GBempty context, 128 generatedvsesmart.czVse Smart review: Mac Studio M5 Ultra (80-core GPU, 256 GB), macOS 27.0.1, measured 24 Oct; MLX 0.32 with mlx-lm, point version not stated. 4-bit MLX; the review does not name the repo.
sourced149 tok/s4-bitllama.cpp256 GBempty context, 128 generatedvsesmart.czVse Smart review: Mac Studio M5 Ultra (80-core GPU, 256 GB), macOS 27.0.1, measured 24 Oct; llama.cpp with every layer on the GPU, build not stated. 4-bit; the review does not name the GGUF type.
Provenance key
sourced
Published figure — vendor or press review
estimated
community
Linked individual run — blog, forum or submission

Speed by context depth: Qwen 3 8B on the M5 Ultra

Measured figures only, each from its source’s own session; the index does not estimate long-context or prefill speed, so a missing cell is —. Each @ column is plain decode with that many tokens already in the prompt, from the same session as the row’s short-context figure: decode slows as the KV cache it reads grows. Prefill is prompt processing, the tokens per second a runtime reads a prompt at; with long agent contexts it sets how long you wait before the first token.

RunDecode (row)@16KPrefill
MLX · 4-bit · 256 GB Mac · vsesmart.czsourced176 tok/s
empty context, 128 generated
sourced121 tok/ssourced6360 tok/s
at 4K
llama.cpp · 4-bit · 256 GB Mac · vsesmart.czsourced149 tok/s
empty context, 128 generated
—sourced4205 tok/s
at 4K

Setup: run Qwen 3 8B 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 2026-11-14:

ollama run qwen3:8b

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: Ollama page → · Hugging Face model card →

The 149 tok/s figure above came from llama.cpp 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.

Qwen 3 8B on other Apple Silicon chips

ChipSpeedQuantEngineBandwidth
M5 Maxestimated91 tok/sQ4_K_MOllama614 GB/s
M5 Prosourced54 tok/s4-bitllama.cpp307 GB/s
M4 Proestimated44 tok/sQ4_K_MOllama273 GB/s
M2estimated17 tok/sQ4_K_MOllama100 GB/s
M1estimated11 tok/sQ4_K_MOllama68 GB/s

For the estimated figures here, the bandwidth formula’s typical error for dense models is ±9%, against 75 measured runs of 20 models (how it is computed).

About Qwen 3 8B

Alibaba's 8B model, ~91 tok/s on M5 Max. Apache 2.0 license, LLMCheck capability 15/50.

Apache 2.0 licence · 8B · 128K context · LLMCheck capability 15/50 · see its leaderboard rank →

Memory by context length

Estimated, not measured: the index’s minimum for Qwen 3 8B (6 GB — the weights at the quant that fits, plus ~2 GB) plus the KV cache for that many tokens, computed from its config.json. GQA on all 36 layers (8 kv heads x 128). A 16-bit KV cache is the default in llama.cpp, Ollama, LM Studio and mlx-lm; a q8 cache halves the KV part. Coding agents hold 64K tokens or more. Runtimes also add a buffer that grows with context, so leave a few GB spare. macOS lets the GPU use only part of a Mac’s memory by default (figures for an M5 Ultra Mac, which runs macOS 26 or later): “fits” means within the GPU memory a Mac of that tier can use with the limit raised, a tier tagged raise GPU limit fits only after sudo sysctl iogpu.wired_limit_mb raises it (the tooltip has the exact command), and the last column is the smallest Mac memory size whose default limit holds it (the GPU memory limit).

ContextKV cache (16-bit)RAM neededRAM, q8 KVSmallest tier that fitsWithin the default GPU limit
8K~1.2 GB~6.6 GB~6.3 GB16 GB Macsame
32K~4.8 GB~10.2 GB~8.1 GB16 GB Macsame
64K (coding-agent context)~9.7 GB~15.1 GB~10.5 GB24 GB Macsame
128K~19.3 GB~24.7 GB~15.4 GB32 GB Macsame

System requirements

Macs with the M5 Ultra

🛒 Get a Mac that runs Qwen 3 8B

Qwen 3 8B needs about 6 GB of unified memory. These Apple Silicon Macs have enough to run it.

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 Qwen 3 8B stacks up against other models on your specific Mac hardware.

Open Compare Tool Full Leaderboard