Bonsai 27B runs at 66 tok/s (66.4 as recorded) on an M5 Max, a published figure reported by huggingface.co (llama.cpp, Q1_0_g128, tg128). It needs about 6 GB of unified memory. It is Prism ML's 27B dense model, released under the Apache 2.0 licence.
Want to run Bonsai 27B faster — or step up to bigger models? Find your Mac →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.
| Metric | Value |
|---|---|
| Tokens per second | sourced66 tok/s |
| Provenance | published figure · huggingface.co |
| Quantization | Q1_0_g128 |
| Engine | llama.cpp |
| Context | tg128 |
| Figure dated | 2026-07 |
| Minimum RAM | 6 GB |
| Parameters | 27B |
| Chip bandwidth | 614 GB/s |
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.
| Speed | Quant | Engine | Mac RAM | Context | Source | Note |
|---|---|---|---|---|---|---|
| sourced66.4 tok/s | Q1_0_g128 | llama.cpp | — | tg128 | huggingface.co | Vendor model card: PrismML llama.cpp fork (Metal), llama-bench; 3.9 GB. |
Ollama — the index has not verified an Ollama tag for Bonsai 27B, so it doesn’t print one (a guessed tag fails with “pull model manifest: file does not exist”). Search the Ollama library →
MLX — pick a build from Hugging Face’s MLX listings, then run it with its repo id:
Downloads: Search the Ollama library → · model card →
The 66 tok/s figure above came from llama.cpp at Q1_0_g128; 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.
Prism ML's 1-bit 27B (Q1_0_g128, 3.9 GB), Apache 2.0, so it runs on an 8 GB Mac. 1-bit decoding is kernel-bound, so the index quotes only vendor figures: 66.4 tok/s on an M5 Max, 44.2 on an M5 Pro and 26.0 on an M4 Pro (PrismML llama.cpp fork, tg128). Its ternary successor is Ternary Bonsai 2 27B.
Apache 2.0 licence · 27B · 262K context · LLMCheck capability 33/50 · see its leaderboard rank →
Bonsai 27B needs about 6 GB of unified memory. These current Apple Silicon Macs have the headroom to run it comfortably:
As an Amazon Associate, LLMCheck earns from qualifying purchases. These affiliate links cost you nothing extra and help keep the LLMCheck index free.
See how Bonsai 27B stacks up against other models on your specific Mac hardware.
Open Compare Tool Full Leaderboard