Yes — Qwen 4 (32B) runs at 33 tok/s on M3 Pro with 18 GB RAM using Q4_K_M quantization via Ollama. First token latency is 1.1s. A capable open-source LLM with 32B parameters.
LLMCheck measured Qwen 4 on M3 Pro using the standard methodology: Q4_K_M quantization, 256-token input, 512-token output, 3 runs averaged on a freshly-booted system.
| Metric | Value |
|---|---|
| Tokens per second | 33 tok/s |
| Time to first token | 1.1s |
| Quantization | Q4_K_M |
| Minimum RAM | 18 GB |
| Recommended engine | Ollama |
| Parameters | 32B |
| Benchmark date | 2026-06 |
Q4_K_M 32B Ollama M3 Pro
The recommended engine for Qwen 4 on M3 Pro is Ollama. Install Ollama, then pull the model:
Ollama handles quantization automatically — it will download the Q4_K_M variant (~18 GB) and start an interactive chat session.
| Chip | Speed | First Token | Min RAM | Engine |
|---|---|---|---|---|
| M5 Max | 80 tok/s | 0.4s | 128 GB | MLX |
| M5 Max | 68 tok/s | 0.5s | 64 GB | Ollama |
| M4 Max | 62 tok/s | 0.6s | 128 GB | Ollama |
| M4 Pro | 60 tok/s | 0.7s | 24 GB | Ollama |
| M5 Pro | 55 tok/s | 0.6s | 64 GB | MLX |
| M3 Max | 47 tok/s | 0.8s | 64 GB | Ollama |
To run Qwen 4 on M3 Pro you need:
See how Qwen 4 stacks up against other models on your specific Mac hardware.
Open Compare Tool Full Leaderboard