Yes: Bonsai 27B runs on 53 of 54 Macs in the LLMCheck index. It needs about 6 GB at a 4K context, 11 GB at 64K. Smallest current Mac: the Mac mini M6 16 GB; cheapest at 15+ tok/s: the Mac mini M5 Pro 24 GB ($1,699; 44 tok/s published). Best measured figure: 66.4 tok/s on an M5 Max (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 | Fits (no speed figure) | — | — | 91K | — |
| MacBook Air M5 24 GB | Fits (no speed figure) | — | — | 179K | — |
| MacBook Air M5 32 GB | Fits (no speed figure) | — | — | 256K (full) | — |
| MacBook Air M1 8 GB refurb | Over GPU limit why ↓ | — | — | — | — |
| MacBook Air M4 16 GB used | Fits (no speed figure) | — | — | 73K | — |
| MacBook Air M2 16 GB refurb | Fits (no speed figure) | — | — | 73K | — |
| MacBook Air M3 16 GB refurb | Fits (no speed figure) | — | — | 73K | — |
| MacBook Air M1 16 GB refurb | Fits (no speed figure) | — | — | 73K | — |
| MacBook Air M4 24 GB used | Fits (no speed figure) | — | — | 153K | — |
| MacBook Air M2 24 GB refurb | Fits (no speed figure) | — | — | 153K | — |
| MacBook Air M3 24 GB refurb | Fits (no speed figure) | — | — | 153K | — |
| MacBook Pro | |||||
| MacBook Pro M6 16 GB | Fits (no speed figure) | — | — | 91K | — |
| MacBook Pro M5 Pro 24 GB | Runs | Q1_0_g128 | sourced44 tok/s | 179K | M5 Pro page |
| MacBook Pro M6 24 GB | Fits (no speed figure) | — | — | 179K | — |
| MacBook Pro M6 32 GB | Fits (no speed figure) | — | — | 256K (full) | — |
| MacBook Pro M5 Pro 48 GB | Runs | Q1_0_g128 | sourced44 tok/s | 256K (full) | M5 Pro page |
| MacBook Pro M5 Max 64 GB | Runs | Q1_0_g128 | sourced66 tok/s | 256K (full) | M5 Max page |
| MacBook Pro M5 Pro 64 GB | Runs | Q1_0_g128 | sourced44 tok/s | 256K (full) | M5 Pro page |
| MacBook Pro M5 Max 128 GB | Runs | Q1_0_g128 | sourced66 tok/s | 256K (full) | M5 Max page |
| MacBook Pro M3 Max 64 GB refurb | Fits (no speed figure) | — | — | 256K (full) | — |
| MacBook Pro M3 Max 128 GB refurb | Fits (no speed figure) | — | — | 256K (full) | — |
| Mac mini | |||||
| Mac mini M6 16 GB | Fits (no speed figure) | — | — | 91K | — |
| Mac mini M5 Pro 24 GB | Runs | Q1_0_g128 | sourced44 tok/s | 179K | M5 Pro page |
| Mac mini M6 24 GB | Fits (no speed figure) | — | — | 179K | — |
| Mac mini M6 32 GB | Fits (no speed figure) | — | — | 256K (full) | — |
| Mac mini M5 Pro 48 GB | Runs | Q1_0_g128 | sourced44 tok/s | 256K (full) | M5 Pro page |
| Mac mini M5 Pro 64 GB | Runs | Q1_0_g128 | sourced44 tok/s | 256K (full) | M5 Pro page |
| Mac mini M2 Pro 16 GB refurb | Fits (no speed figure) | — | — | 73K | — |
| Mac mini M4 16 GB used | Fits (no speed figure) | — | — | 73K | — |
| Mac mini M4 Pro 24 GB used | Runs | Q1_0_g128 | sourced26 tok/s | 153K | M4 Pro page |
| Mac mini M4 24 GB used | Fits (no speed figure) | — | — | 153K | — |
| Mac mini M2 Pro 32 GB refurb | Fits (no speed figure) | — | — | 232K | — |
| Mac mini M4 Pro 48 GB used | Runs | Q1_0_g128 | sourced26 tok/s | 256K (full) | M4 Pro page |
| Mac mini M4 Pro 64 GB used | Runs | Q1_0_g128 | sourced26 tok/s | 256K (full) | M4 Pro page |
| iMac | |||||
| iMac M6 16 GB | Fits (no speed figure) | — | — | 91K | — |
| iMac M6 24 GB | Fits (no speed figure) | — | — | 179K | — |
| iMac M6 32 GB | Fits (no speed figure) | — | — | 256K (full) | — |
| Mac Studio | |||||
| Mac Studio M5 Max 36 GB | Fits (no speed figure) | — | — | 256K (full) | M5 Max page |
| Mac Studio M5 Max 48 GB | Runs | Q1_0_g128 | sourced66 tok/s | 256K (full) | M5 Max page |
| Mac Studio M5 Max 64 GB | Runs | Q1_0_g128 | sourced66 tok/s | 256K (full) | M5 Max page |
| Mac Studio M5 Ultra 96 GB | Fits (no speed figure) | — | — | 256K (full) | — |
| Mac Studio M5 Max 128 GB | Runs | Q1_0_g128 | sourced66 tok/s | 256K (full) | M5 Max page |
| Mac Studio M5 Ultra 256 GB | Fits (no speed figure) | — | — | 256K (full) | — |
| Mac Studio M5 Ultra 512 GB | Fits (no speed figure) | — | — | 256K (full) | — |
| Mac Studio M2 Max 32 GB refurb | Fits (no speed figure) | — | — | 232K | — |
| Mac Studio M1 Ultra 64 GB refurb | Fits (no speed figure) | — | — | 256K (full) | — |
| Mac Studio M4 Max 64 GB used | Fits (no speed figure) | — | — | 256K (full) | — |
| Mac Studio M2 Max 64 GB refurb | Fits (no speed figure) | — | — | 256K (full) | — |
| Mac Studio M1 Ultra 128 GB refurb | Fits (no speed figure) | — | — | 256K (full) | — |
| Mac Studio M2 Ultra 128 GB refurb | Fits (no speed figure) | — | — | 256K (full) | — |
| Mac Studio M4 Max 128 GB used | Fits (no speed figure) | — | — | 256K (full) | — |
| Mac Studio M2 Ultra 192 GB refurb | Fits (no speed figure) | — | — | 256K (full) | — |
| Mac Studio M3 Ultra 256 GB used | Fits (no speed figure) | — | — | 256K (full) | — |
| Mac Studio M3 Ultra 512 GB refurb | Fits (no speed figure) | — | — | 256K (full) | — |
Context memory is computed from its config.json (config.json read from the same org's prism-ml/Bonsai-27B-unpacked (same architecture; the card repo has none). hybrid: full attention on 16 of 64 layers; Gated DeltaNet layers keep a constant state): about 4.3 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 wire only part of unified memory by default (Metal’s recommended working set; M5 and M6
Macs run macOS 26 or later, whose default is higher), and Bonsai 27B needs about 6 GB. On the Macs marked
Raise GPU limit it fits in RAM but not inside that default. The default
is a performance threshold, not a hard wall: past it, the runtime moves part of it to the CPU or shortens the context (llama.cpp, Ollama), runs it slowly (MLX), or stops with an out-of-memory error. One command raises the limit until the next restart
(sudo sysctl iogpu.wired_limit_mb=0 restores the default sooner); it leaves macOS a reserve, and other
apps get less memory while the model is loaded. On the Macs marked Over GPU limit even that raise falls short: the model fits the Mac’s RAM on paper, but a smaller quant or a bigger Mac is the realistic route.
Bonsai 27B ships as a sub-2-bit / ternary build, which decodes through its own kernels; the GGUF quant ladder (Q2_K…F16) does not describe it. The index needs about 6 GB for it at a short context.
Every published (filled square) and community (hollow circle) run the index holds for Bonsai 27B: 3 published, 0 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 Max | llama.cpp | Q1_0_g128 | sourced66.4 tok/s | tg128 | 2026-07 | huggingface.co |
| M5 Pro | llama.cpp | Q1_0_g128 | sourced44.2 tok/s | tg128 | 2026-07 | huggingface.co |
| M4 Pro | llama.cpp | Q1_0_g128 | sourced26 tok/s | tg128 | 2026-07 | huggingface.co |
The index has no verified Ollama tag for Bonsai 27B. Start from its listing and copy the exact name of a build that fits: Search the Ollama library → · model card →
GGUF builds run in LM Studio or llama.cpp; MLX builds run in LM Studio or mlx-lm.
On the Mac mini M6 (16 GB), the smallest current Mac that runs it inside the default GPU limit, the index figure is not available; keep the context at or under 91K there, or use a q8 KV cache.
Step-by-step: the Ollama install guide.
Mac mini M6 (16 GB) — smallest current Mac that runs it, $899 new (Apple list price, checked 2026-11-14). Mac mini M5 Pro (24 GB) — cheapest at 15+ tok/s, $1,699 new (Apple list price, checked 2026-11-14), 44 tok/s (published figure).
As an Amazon Associate, LLMCheck earns from qualifying purchases. Affiliate links cost you nothing extra and never influence the verdicts.
Yes. Bonsai 27B needs about 6 GB of unified memory at a short (4K-token) context, so it runs on 53 of the 54 Mac configurations in the LLMCheck index. On 1 more it fits the RAM, but not even a safely raised GPU limit. The smallest Mac Apple sells new that runs it is the Mac mini M6 (16 GB).
About 6 GB at a 4K context: the weights plus about 2 GB for the runtime. Context adds KV cache: about 8 GB in total at 32K and 11 GB at 64K with the default 16-bit cache (about 9 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 66 tok/s (published figure) at Q1_0_g128. The best measured figure is 66.4 tok/s on an M5 Max (published figure, Q1_0_g128). 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 M5 Pro (24 GB), $1,699 new (Apple list price, checked 2026-11-14), is the cheapest that reaches 15 tok/s with Bonsai 27B, at 44 tok/s (published figure).
On MacBook Air M1 8 GB (used) even a safely raised limit is short: expect heavy swapping, a CPU split or an out-of-memory error. From 16 GB up, every Mac Apple sells new holds it inside the default limit.
Bonsai 27B ships as a sub-2-bit / ternary build with its own kernels, so the usual GGUF ladder (Q2_K to F16) does not apply: use the published build.
Its own window is 256K tokens. At macOS’s default GPU limit, the Mac mini M6 (16 GB) holds about 91K. The full window needs about 23 GB, which a 32 GB Mac holds at the default limit. A q8 KV cache roughly doubles these lengths.
LLMCheck. (2026). Can I run Bonsai 27B on a Mac? llmcheck.net. https://llmcheck.net/can-i-run/bonsai-27b/ (updated ). CC BY 4.0.