M4 Pro (48 GB) — benchmark record
M4 Pro (48 GB) local LLM benchmarks on Apple Silicon Mac with 3 published rows across 3 models. Peak published speed is 55.0 tok/s on Qwen 3 30B-A3B. Largest published fit uses 18.5 GB. Published runtimes include LM Studio, MLX.
Quick take
Best published speed here is 55.0 tok/s on Qwen 3 30B-A3B at 8bit. Largest published memory footprint is Devstral Small 1.1 at 6bit, using 18.5 GB. Longest published context on this page is 131k. This page is evidence, not the full buying answer.
Based on 3 external benchmarks; no lab runs yet.
Published runtimes: LM Studio, MLX.
Current published coverage
Published model coverage includes Qwen 3 30B-A3B, Devstral Small 1.1, Qwen3.5-27B. Published runtime coverage on this chip includes LM Studio, MLX. Fastest published row is 55.0 tok/s on Qwen 3 30B-A3B at 8bit. Largest published footprint on this page is 18.5 GB for Devstral Small 1.1. Longest published context is 131k.
Macs shipping with the M4 Pro (48 GB)
This chip is part of a family. Compare all M4 Pro RAM variants →
Raw benchmark rows for M4 Pro (48 GB)
Rows sorted by avg tok/s descending. Click source badge to see original measurement page.
| Model | Quant | Avg tok/s | Runtime | Source |
|---|---|---|---|---|
| Qwen 3 30B-A3B | 8bit | 55.0 tok/s | MLX | ref |
| Devstral Small 1.1 | 6bit | 12.9 tok/s | LM Studio | ref |
| Qwen3.5-27B | 8bit | 8.5 tok/s | MLX | ref |
Models tested on this chip
Next step
Use Rankings when you need the best overall answer for a Mac, and keep this page open when you need the evidence behind that recommendation.
If you are comparing local Apple Silicon against rented cloud hardware, use AI Data Center Index for current GPU rental context.
Data
benchmarks.json — full dataset · chips.json — chip summaries · benchmarks.csv — CSV export
Data from in-house lab measurements plus community-published benchmarks. See all chip families →