RX 7900 XTX 24GBBudget pick
960 GB/s (+560)
C
Raises estimated decode speed by about 136%.112 tok/s decode
Raises estimated decode speed by about 136%.
~$999 MSRP
Can DeepSeek R1 0528 Qwen3 8B run on MacBook Pro M2 Max 32GB?
YES — Runs Great
C49Usable
○Estimated from fit model
DeepSeek R1 0528 Qwen3 8B needs ~10.2 GB VRAM. MacBook Pro M2 Max 32GB has 23.0 GB. With Q4_K_M quantization, expect ~48 tok/s.
Runtime: llama.cppCapacity: RoomyBandwidth: LowStack: StandardBottleneck: Balanced
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) — 10.2 GB, 47.5 tok/s, Runs well
10.2 GB required23.0 GB available
44% VRAM used
Fit status
Runs well
Decode
47.5 tok/s
TTFT
4072 ms
Safe context
236K
Memory
10.2 GB / 23.0 GB
Memory breakdown
Weights4.9 GB
KV Cache0.9 GB
Runtime0.9 GB
Headroom3.5 GB
See how fast it feels
See how fast it feelsDeepSeek R1 0528 Qwen3 8B on MacBook Pro M2 Max 32GB
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: 47.5 tok/s decode · 4.1s TTFT (warm) · 119 tok/s prefill
What limits this setup
This setup is broadly balanced for this model.
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
Performance by workload
| Workload | Grade | Fit | Decode | TTFT | Context |
|---|---|---|---|---|---|
| Chat | C | Runs well | 47.5 tok/s | 2221 ms | 236K |
| Coding | C | Runs well | 47.5 tok/s | 4072 ms | 236K |
| Agentic Coding | C | Runs well | 47.5 tok/s | 5923 ms | 236K |
| Reasoning | C | Runs well | 47.5 tok/s | 4813 ms | 236K |
| RAG | C | Runs well | 47.5 tok/s | 7404 ms | 236K |
Quantization options
How DeepSeek R1 0528 Qwen3 8B (8B params) fits at each quantization level on MacBook Pro M2 Max 32GB (23.0 GB usable).
| Quant | Bits | VRAM | Quality | Fit |
|---|---|---|---|---|
Q2_K | 2 | 3.1 GB | Low | C45 |
Q3_K_S | 3 | 3.9 GB | Low | C45 |
NVFP4 | 4 | 4.5 GB | Medium | C46 |
Q4_K_M | 4 | 4.9 GB | Medium | C46 |
Q5_K_M | 5 | 5.8 GB | High | C46 |
Q6_K | 6 | 6.6 GB | High | C47 |
Q8_0 | 8 | 8.6 GB | Very High | C48 |
F16Best for your GPU | 16 | 16.4 GB | Maximum | C50 |
Get started
Copy-paste commands to run DeepSeek R1 0528 Qwen3 8B on your machine.
Run
lms load hf-maziyarpanahi--deepseek-r1-0528-qwen3-8b-gguf && lms server startUpgrade options
