MacBook Pro M4 32GBBudget pick
32 GB Unified (+8)120 GB/s (+20)
B
Adds memory headroom for longer context windows and future model growth.9.6 tok/s decode
Adds memory headroom for longer context windows and future model growth.
~$799 MSRP
Can Phi 3 Medium 14B run on Mac mini M2 24GB?
YES — Tight Fit
B57Good
○Estimated from fit model
Phi 3 Medium 14B needs ~15.1 GB VRAM. Mac mini M2 24GB has 17.3 GB. With Q4_K_M quantization, expect ~8 tok/s.
Runtime: llama.cppCapacity: TightBandwidth: Very lowStack: StandardBottleneck: Memory bandwidth
Share:
Operating mode
Choose the run profile you care about
Interactive favors responsiveness, while light API and scale-out lean harder on serving readiness. The fit stays the same, but the recommendation lens changes.
Current mode
Balanced
Balanced for general local use. Keeps the ranking neutral across personal and serving workflows.
Select quantization to explore
Q4_K_M (Medium quality) — 15.1 GB, 8.2 tok/s, Tight fit
15.1 GB required17.3 GB available
87% VRAM used
Fit status
Tight fit
Decode
8.2 tok/s
TTFT
23661 ms
Safe context
28K
Memory
15.1 GB / 17.3 GB
Memory breakdown
Weights8.5 GB
KV Cache3.1 GB
Runtime0.9 GB
Headroom2.6 GB
See how fast it feels
See how fast it feelsPhi 3 Medium 14B on Mac mini M2 24GB
1st promptCold start — includes initialization
>Solve: If a train travels at 60 mph for 2.5 hours, how far does it go?
Let me work through this step by step.
Given information:
- Speed = 60 mph (miles per hour)
- Time = 2.5 hours
Formula: Distance = Speed x Time
Calculation:
Distance = 60 mph x 2.5 h
= 150 miles
The train travels 150 miles.
Double-check: at 60 mph the train covers 60 miles each hour. In half an hour it covers 30 miles. So in 2.5 hours: 60 + 60 + 30 = 150 miles. Confirmed.
2nd promptWarm — model ready, no init overhead
>What if the train stops for 20 minutes halfway?
The stop doesn't change the total distance covered while moving, but it affects total trip time.
Breakdown:
- First half: 1.25 hours at 60 mph = 75 miles
- Stop: 20 minutes (no distance)
- Second half: 1.25 hours at 60 mph = 75 miles
Total distance: 150 miles (same)
Total time: 2.5 h + 0.33 h = 2.83 hours
Average speed: 150 / 2.83 = 53 mph
The distance stays the same but average speed drops to 53 mph because of the stop.
Estimated: 8.2 tok/s decode · 23.7s TTFT (warm) · 21 tok/s prefill
What limits this setup
The model fits in shared memory, but shared-memory bandwidth is now the real limiter.
Fit does not mean dedicated-VRAM speed
Unified or shared memory can make a model technically fit, but sustained tokens per second may still trail a discrete high-bandwidth GPU with less total memory.
Shared-memory contention still exists
The OS, browser, and inference runtime all compete for the same physical memory pool, so real-world headroom is less forgiving than raw capacity suggests.
Best improvement path
Prioritize bandwidth, not only capacity
If this workload feels slow, the next useful step is often a GPU tier with materially faster memory bandwidth rather than only a small bump in capacity.
Performance by workload
| Workload | Grade | Fit | Decode | TTFT | Context |
|---|---|---|---|---|---|
| Chat | B | Runs well | 7.6 tok/s | 13874 ms | 28K |
| Coding | B | Tight fit | 7.6 tok/s | 25436 ms | 28K |
| Agentic Coding | B | Runs with offload | 6.9 tok/s | 40518 ms | 28K |
| Reasoning | B | Tight fit | 7.6 tok/s | 30061 ms | 28K |
| RAG | B | Runs with offload | 6.9 tok/s | 50648 ms | 28K |
Quantization options
How Phi 3 Medium 14B (14B params) fits at each quantization level on Mac mini M2 24GB (17.3 GB usable).
| Quant | Bits | VRAM | Quality | Fit |
|---|---|---|---|---|
Q2_K | 2 | 5.5 GB | Low | B60 |
Q3_K_S | 3 | 6.9 GB | Low | B61 |
NVFP4 | 4 |
Get started
Copy-paste commands to run Phi 3 Medium 14B on your machine.
Run
ollama run phi3:mediumUpgrade options
Hardware that runs Phi 3 Medium 14B well
Mac mini M4 32GBBest value
32 GB Unified (+8)120 GB/s (+20)
B
Adds memory headroom for longer context windows and future model growth.9.6 tok/s decode
Adds memory headroom for longer context windows and future model growth.
~$1,099 MSRP
640 GB/s (+540)
B
Raises estimated decode speed by about 666%.62.8 tok/s decode
Raises estimated decode speed by about 666%.
~$2,000 MSRP
Mac Studio M2 Ultra 64GBApple upgrade
64 GB Unified (+40)800 GB/s (+700)
B
Raises estimated decode speed by about 612%.58.4 tok/s decode
Raises estimated decode speed by about 612%.
Adds memory headroom for longer context windows and future model growth.
~$3,999 MSRP