Run Qwen 3.5 4B on M6

The LLMCheck index estimates Qwen 3.5 4B at ~54 tok/s on an M6 at Q4_K_M, a figure from its published bandwidth model, not a measured run. It needs about 3 GB of unified memory. It is Alibaba's 4B dense model, released under the Apache 2.0 licence.

Speed
estimated~54
tok/s
First Token
—
seconds
RAM Needed
3
GB minimum
Engine
—
estimate
Want to run Qwen 3.5 4B faster — or step up to bigger models? Find your Mac →

The figure

This is an LLMCheck index estimate from its published two-term bandwidth model, not a measured run; the index labels it so everywhere it appears. The bandwidth formula’s typical error for dense models is ±9%, against 75 measured runs of 20 models (how it is computed). Own this Mac? Submit a real run →

MetricValue
Tokens per secondestimated~54 tok/s
Provenanceestimate · how estimates are made
QuantizationQ4_K_M
Minimum RAM3 GB
Parameters4B
Chip bandwidth170 GB/s

Every figure for Qwen 3.5 4B on the M6

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 (3 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
community52.4 tok/s4-bitMLX16 GB1komlx.aiThe 16 GB M6 is the binned 153 GB/s part, so this run never anchors or sets the reference. 12-core GPU, oMLX v0.7.3, MTP/DFlash off; Qwen3.5-4B-MLX-4bit; 3.4 GB peak. 50.4 tok/s at 4k.; binned GPU
Provenance key
sourced
Published figure — vendor or press review
estimated
community
Linked individual run — blog, forum or submission

Speed by context depth: Qwen 3.5 4B on the M6

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)@4KPrefillTTFT
MLX · 4-bit · 16 GB Mac · omlx.aicommunity52.4 tok/s
1k
community50.4 tok/scommunity1813 tok/s
at 4K
community0.622 s

Setup: run Qwen 3.5 4B 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 qwen3.5:4b

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 →

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

Qwen 3.5 4B on other Apple Silicon chips

ChipSpeedQuantEngineBandwidth
M5 Maxestimated161 tok/sQ4_K_MMLX614 GB/s
M5 Prosourced66 tok/sQ4_K_Mllama.cpp307 GB/s
M4estimated39 tok/sQ4_K_MOllama120 GB/s

About Qwen 3.5 4B

Alibaba's 4B model, ~161 tok/s on M5 Max. Apache 2.0 license, LLMCheck capability 12/50.

Apache 2.0 licence · 4B · 262K context · LLMCheck capability 12/50 · see its leaderboard rank →

Memory by context length

Estimated, not measured: the index’s minimum for Qwen 3.5 4B (3 GB — the weights at the quant that fits, plus ~2 GB) plus the KV cache for that many tokens, computed from its config.json. hybrid: full attention on 8 of 32 layers; Gated DeltaNet layers keep a constant state. 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 M6 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~0.3 GB~3.1 GB~3.1 GB8 GB Macsame
32K~1.1 GB~3.9 GB~3.5 GB8 GB Macsame
64K (coding-agent context)~2.1 GB~5.0 GB~4.0 GB8 GB Macsame
128K~4.3 GB~7.2 GB~5.1 GB16 GB Macsame

System requirements

Macs with the M6

🛒 Get a Mac that runs Qwen 3.5 4B

Qwen 3.5 4B needs about 3 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 3.5 4B stacks up against other models on your specific Mac hardware.

Open Compare Tool Full Leaderboard