Yes: Qwen 3.5 4B runs on all 54 Macs in the LLMCheck index. It needs about 3 GB at a 4K context, 6 GB at 64K. The smallest and cheapest current Mac that runs it at 15+ tok/s is the Mac mini M6 16 GB ($899; 52 tok/s community-measured). Best measured figure: 65.82 tok/s on an M5 Pro (published).
One row per Mac configuration, grouped by product line. The build is the one each speed is for; the context column is the longest prompt that fits inside macOS’s default GPU memory limit with a 16-bit KV cache (“raised” where the limit has to be raised first; “full” when the whole window fits). Speeds are decode tok/s from the LLMCheck index, each with its provenance mark. For dense models, the formula’s typical error is about 9% (the median gap to 75 measured plain-decode runs; it tends to undershoot (median −6%), and an estimate scaled from the model’s own measured run is usually closer; method).
| Mac | Verdict | Build | Speed | Max context | Model page |
|---|---|---|---|---|---|
| MacBook Air | |||||
| MacBook Air M5 16 GB | Runs | Q4_K_M | estimated49 tok/s | 256K (full) | — |
| MacBook Air M5 24 GB | Runs | Q4_K_M | estimated49 tok/s | 256K (full) | — |
| MacBook Air M5 32 GB | Runs | Q4_K_M | estimated49 tok/s | 256K (full) | — |
| MacBook Air M1 8 GB refurb | Runs | Q4_K_M | estimated23 tok/s | 73K | — |
| MacBook Air M4 16 GB used | Runs | Q4_K_M | estimated39 tok/s | 232K | — |
| MacBook Air M2 16 GB refurb | Runs | Q4_K_M | estimated33 tok/s | 232K | — |
| MacBook Air M3 16 GB refurb | Runs | Q4_K_M | estimated33 tok/s | 232K | — |
| MacBook Air M1 16 GB refurb | Runs | Q4_K_M | estimated23 tok/s | 232K | — |
| MacBook Air M4 24 GB used | Runs | Q4_K_M | estimated39 tok/s | 256K (full) | — |
| MacBook Air M2 24 GB refurb | Runs | Q4_K_M | estimated33 tok/s | 256K (full) | — |
| MacBook Air M3 24 GB refurb | Runs | Q4_K_M | estimated33 tok/s | 256K (full) | — |
| MacBook Pro | |||||
| MacBook Pro M6 16 GB | Runs | 4-bit | community52 tok/s | 256K (full) | M6 page |
| MacBook Pro M5 Pro 24 GB | Runs | Q4_K_M | sourced66 tok/s | 256K (full) | M5 Pro page |
| MacBook Pro M6 24 GB | Runs | Q4_K_M | estimated54 tok/s | 256K (full) | M6 page |
| MacBook Pro M6 32 GB | Runs | Q4_K_M | estimated54 tok/s | 256K (full) | M6 page |
| MacBook Pro M5 Pro 48 GB | Runs | Q4_K_M | sourced66 tok/s | 256K (full) | M5 Pro page |
| MacBook Pro M5 Max 64 GB | Runs | Q4_K_M | estimated161 tok/s | 256K (full) | M5 Max page |
| MacBook Pro M5 Pro 64 GB | Runs | Q4_K_M | sourced66 tok/s | 256K (full) | M5 Pro page |
| MacBook Pro M5 Max 128 GB | Runs | Q4_K_M | estimated161 tok/s | 256K (full) | M5 Max page |
| MacBook Pro M3 Max 64 GB refurb | Runs | Q4_K_M | estimated83 tok/s | 256K (full) | — |
| MacBook Pro M3 Max 128 GB refurb | Runs | Q4_K_M | estimated83 tok/s | 256K (full) | — |
| Mac mini | |||||
| Mac mini M6 16 GB | Runs | 4-bit | community52 tok/s | 256K (full) | M6 page |
| Mac mini M5 Pro 24 GB | Runs | Q4_K_M | sourced66 tok/s | 256K (full) | M5 Pro page |
| Mac mini M6 24 GB | Runs | Q4_K_M | estimated54 tok/s | 256K (full) | M6 page |
| Mac mini M6 32 GB | Runs | Q4_K_M | estimated54 tok/s | 256K (full) | M6 page |
| Mac mini M5 Pro 48 GB | Runs | Q4_K_M | sourced66 tok/s | 256K (full) | M5 Pro page |
| Mac mini M5 Pro 64 GB | Runs | Q4_K_M | sourced66 tok/s | 256K (full) | M5 Pro page |
| Mac mini M2 Pro 16 GB refurb | Runs | Q4_K_M | estimated59 tok/s | 232K | — |
| Mac mini M4 16 GB used | Runs | Q4_K_M | estimated39 tok/s | 232K | — |
| Mac mini M4 Pro 24 GB used | Runs | Q4_K_M | estimated59 tok/s | 256K (full) | — |
| Mac mini M4 24 GB used | Runs | Q4_K_M | estimated39 tok/s | 256K (full) | — |
| Mac mini M2 Pro 32 GB refurb | Runs | Q4_K_M | estimated59 tok/s | 256K (full) | — |
| Mac mini M4 Pro 48 GB used | Runs | Q4_K_M | estimated59 tok/s | 256K (full) | — |
| Mac mini M4 Pro 64 GB used | Runs | Q4_K_M | estimated59 tok/s | 256K (full) | — |
| iMac | |||||
| iMac M6 16 GB | Runs | 4-bit | community52 tok/s | 256K (full) | M6 page |
| iMac M6 24 GB | Runs | Q4_K_M | estimated54 tok/s | 256K (full) | M6 page |
| iMac M6 32 GB | Runs | Q4_K_M | estimated54 tok/s | 256K (full) | M6 page |
| Mac Studio | |||||
| Mac Studio M5 Max 36 GB | Runs | Q4_K_M | estimated94 tok/s | 256K (full) | M5 Max page |
| Mac Studio M5 Max 48 GB | Runs | Q4_K_M | estimated161 tok/s | 256K (full) | M5 Max page |
| Mac Studio M5 Max 64 GB | Runs | Q4_K_M | estimated161 tok/s | 256K (full) | M5 Max page |
| Mac Studio M5 Ultra 96 GB | Runs | Q4_K_M | estimated261 tok/s | 256K (full) | — |
| Mac Studio M5 Max 128 GB | Runs | Q4_K_M | estimated161 tok/s | 256K (full) | M5 Max page |
| Mac Studio M5 Ultra 256 GB | Runs | Q4_K_M | estimated261 tok/s | 256K (full) | — |
| Mac Studio M5 Ultra 512 GB | Runs | Q4_K_M | estimated261 tok/s | 256K (full) | — |
| Mac Studio M2 Max 32 GB refurb | Runs | Q4_K_M | estimated83 tok/s | 256K (full) | — |
| Mac Studio M1 Ultra 64 GB refurb | Runs | Q4_K_M | estimated196 tok/s | 256K (full) | — |
| Mac Studio M4 Max 64 GB used | Runs | Q4_K_M | estimated147 tok/s | 256K (full) | — |
| Mac Studio M2 Max 64 GB refurb | Runs | Q4_K_M | estimated83 tok/s | 256K (full) | — |
| Mac Studio M1 Ultra 128 GB refurb | Runs | Q4_K_M | estimated196 tok/s | 256K (full) | — |
| Mac Studio M2 Ultra 128 GB refurb | Runs | Q4_K_M | estimated196 tok/s | 256K (full) | — |
| Mac Studio M4 Max 128 GB used | Runs | Q4_K_M | estimated147 tok/s | 256K (full) | — |
| Mac Studio M2 Ultra 192 GB refurb | Runs | Q4_K_M | estimated196 tok/s | 256K (full) | — |
| Mac Studio M3 Ultra 256 GB used | Runs | Q4_K_M | estimated199 tok/s | 256K (full) | — |
| Mac Studio M3 Ultra 512 GB refurb | Runs | Q4_K_M | estimated199 tok/s | 256K (full) | — |
Context memory is computed from its config.json (hybrid: full attention on 8 of 32 layers; Gated DeltaNet layers keep a constant state): about 2.1 GB of 16-bit KV cache at 64K tokens. A q8 cache (llama.cpp -ctk q8_0 -ctv q8_0, Ollama OLLAMA_KV_CACHE_TYPE=q8_0) halves that part. Runtimes also grow a compute buffer with context, so a fit within about 1 GB of the limit is optimistic.
macOS lets the GPU use only part of unified memory by default. Qwen 3.5 4B fits inside that default limit on every Mac it fits at all, so no Mac in the table needs it raised at a short context. Long contexts may still need it: the context column already uses each Mac’s default limit.
Weights are 4B parameters at each format’s effective size (the index’s GB-per-billion table), plus about 2 GB for the runtime and a short context; longer contexts add the KV cache computed from the model’s config.json at 16 bits. The first row is the catalog’s own figure (about 3 GB at a 4K context), the one every verdict on this page uses. “Smallest Mac” is the smallest memory size among the Macs Apple sells new that holds it at a 4K context inside that Mac’s default GPU limit. Estimates, not measurements.
| Build | Weights | RAM at 4K | at 32K | at 64K | Smallest Mac |
|---|---|---|---|---|---|
| As catalogued | — | 3 GB | 4 GB | 5 GB | 16 GB |
| Q2_K | 1.4 GB | 3 GB | 4 GB | 5 GB | 16 GB |
| Q3_K_M | 2.0 GB | 4 GB | 5 GB | 6 GB | 16 GB |
| Q4_K_M | 2.3 GB | 4 GB | 5 GB | 6 GB | 16 GB |
| Q5_K_M | 2.8 GB | 5 GB | 6 GB | 7 GB | 16 GB |
| Q6_K | 3.3 GB | 5 GB | 6 GB | 7 GB | 16 GB |
| Q8_0 | 4.2 GB | 6 GB | 7 GB | 8 GB | 16 GB |
| F16 | 8.0 GB | 10 GB | 11 GB | 12 GB | 16 GB |
Every published (filled square) and community (hollow circle) run the index holds for Qwen 3.5 4B: 1 published, 1 community, each linked to its source. LLMCheck runs no tests of its own. One row is one measurement session; on each chip the table above shows the lower median of the eligible runs (plain decode at 5 bits or fewer), and a binned or “not an anchor” run is shown but never used as a reference.
| Chip | Runtime | Quant | Decode | Context | Date | Source |
|---|---|---|---|---|---|---|
| M5 Pro | llama.cpp | Q4_K_M | sourced65.82 tok/s | decode at 4K context depth | 2026-10-21 | huggingface.co |
| M6 | MLX | 4-bit | community52.4 tok/s | 1k | 2026-11-08 | omlx.ai binned |
The quickest start is Ollama. Pull the model by its tag (checked on ollama.com, 2026-11-14; the file is about 4.0 GB):
Prefer a window to a terminal? LM Studio downloads the same model with one click. · model card →
On the Mac mini M6 (16 GB), the smallest current Mac that runs it inside the default GPU limit, the index figure is 52 tok/s (community run) at 4-bit; keep the context at or under 256K there, or use a q8 KV cache. The qwen3.5:4b file fits there.
Step-by-step: the Ollama install guide.
Mac mini M6 (16 GB) — smallest current Mac that runs it and cheapest at 15+ tok/s, $899 new (Apple list price, checked 2026-11-14), 52 tok/s (community run).
As an Amazon Associate, LLMCheck earns from qualifying purchases. Affiliate links cost you nothing extra and never influence the verdicts.
Yes. Qwen 3.5 4B needs about 3 GB of unified memory at a short (4K-token) context, so it runs on 54 of the 54 Mac configurations in the LLMCheck index. The smallest Mac Apple sells new that runs it is the Mac mini M6 (16 GB).
About 3 GB at a 4K context: the weights plus about 2 GB for the runtime. Context adds KV cache: about 4 GB in total at 32K and 6 GB at 64K with the default 16-bit cache (about 5 GB at 64K with a q8 cache). These are estimates from the model’s config.json, not measurements.
On an M5 Max, the chip the leaderboard quotes it on, the index figure is ~161 tok/s (estimated) at Q4_K_M. The best measured figure is 65.82 tok/s on an M5 Pro (published figure, Q4_K_M). Every Mac’s figure is in the table on this page with its provenance: published, community run or estimated.
Among Macs Apple sells new, the Mac mini M6 (16 GB), $899 new (Apple list price, checked 2026-11-14), is the cheapest that reaches 15 tok/s with Qwen 3.5 4B, at 52 tok/s (community run).
No. On all 54 Mac configurations where Qwen 3.5 4B fits, it fits inside that Mac’s default GPU memory limit at a short context; only a long context on the smallest of them can push it over.
Q4_K_M (about 2 GB of weights) is the common 4-bit GGUF choice. Q8_0 (about 4 GB) is closer to full quality and needs a 16 GB Mac at the default GPU limit. Q3_K_M and Q2_K shrink it to about 2 and 1 GB at a quality cost. The index’s reference figure (161 tok/s on an M5 Max) is for its Q4_K_M build.
Its own window is 256K tokens. At macOS’s default GPU limit, the Mac mini M6 (16 GB) holds about 256K (the full window).
LLMCheck. (2026). Can I run Qwen 3.5 4B on a Mac? llmcheck.net. https://llmcheck.net/can-i-run/qwen-35-4b/ (updated ). CC BY 4.0.