Yes — Llama 3.1 405B (405B) runs at 4 tok/s on M5 Ultra with 512 GB RAM using Q4_K_M quantization via MLX. First token latency is Nones. A capable open-source LLM with 405B parameters.
Want to run Llama 3.1 405B faster — or step up to bigger models? Find your Mac →The LLMCheck index estimates Llama 3.1 405B on M5 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 →
| Metric | Value |
|---|---|
| Tokens per second | 4 tok/s |
| Time to first token | Nones |
| Quantization | Q4_K_M |
| Minimum RAM | 512 GB |
| Recommended engine | MLX |
| Parameters | 405B |
| Benchmark date | 2026-08 |
Q4_K_M 405B MLX M5 Ultra
The recommended engine for Llama 3.1 405B on M5 Ultra is MLX. Install with pip and pull the model:
Alternatively, you can use Ollama for a simpler setup:
To run Llama 3.1 405B on M5 Ultra you need:
Llama 3.1 405B needs about 512 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 our benchmarks free.
See how Llama 3.1 405B stacks up against other models on your specific Mac hardware.
Open Compare Tool Full Leaderboard