Run Qwen3.8-27B on M3 Ultra

Yes — Qwen3.8-27B (27.8B) runs at 40 tok/s on M3 Ultra with 96 GB RAM using Q4_K_M quantization via LM Studio. A capable open-source LLM with 27.8B parameters.

Speed
sourced40
tok/s
First Token
seconds
RAM Needed
96
GB minimum
Engine
LM Studio
recommended
Want to run Qwen3.8-27B faster — or step up to bigger models? Find your Mac →

Benchmark Details

The LLMCheck index estimates Qwen3.8-27B on M3 Ultra using our published methodology: Q4_K_M quantization, memory-bandwidth scaling, and cross-referenced third-party benchmarks where available. Figures are transparent estimates — own this config? Submit a real benchmark →

MetricValue
Tokens per second40 tok/s
Time to first token
QuantizationQ4_K_M
Minimum RAM96 GB
Recommended engineLM Studio
Parameters27.8B
Benchmark date2026-09

Q4_K_M 27.8B LM Studio M3 Ultra

Setup Guide: Run Qwen3.8-27B on M3 Ultra

The recommended engine for Qwen3.8-27B on M3 Ultra is LM Studio. Install Ollama, then pull the model:

ollama run qwen38-27b

Ollama handles quantization automatically — it will download the Q4_K_M variant (~96 GB) and start an interactive chat session.

Performance on Other Apple Silicon Chips

ChipSpeedFirst TokenMin RAMEngine
M5 Ultra 55 tok/s 0.5s 256 GB LM Studio
M5 Max 29 tok/s 128 GB MLX
M4 Max 27 tok/s 64 GB MLX
M3 Max 20 tok/s 64 GB MLX
M4 Pro 14 tok/s 48 GB Ollama
M6 8 tok/s 24 GB MLX

System Requirements

To run Qwen3.8-27B on M3 Ultra you need:

🛒 Get a Mac that runs Qwen3.8-27B

Qwen3.8-27B needs about 96 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.

Compare More Models

See how Qwen3.8-27B stacks up against other models on your specific Mac hardware.

Open Compare Tool Full Leaderboard