Can MacBook Pro M3 Max 36 GB run Gemma 4 31B?
36 GB unified · 300 GB/s memory bandwidth · Gemma 4 31B has 31.3B parameters.
Short answer
Yes, comfortable for chat; borderline for agentic coding.
Best quantization for chat: 4-bit (17 GB of weights) on LM Studio. Generation: 11–14 tok/s, leaving about 9.1 GB free for your other apps.
How it rates for each workload
Same model and machine, different jobs. Agents and long documents send far bigger prompts than chat, so they need much more speed and memory.
| Workload | Verdict | Quant | Generation | First reply | Context | |
|---|---|---|---|---|---|---|
| Chatwith Open WebUI | Comfortable | 4-bit | 11–14 tok/s | ≈6 s | 8K | Details |
| Coding questionswith Continue | Acceptable | 4-bit | 10–14 tok/s | ≈40 s | 16K | Details |
| Coding in a repositorywith Aider | Acceptable | 4-bit | 10–13 tok/s | ≈2 min | 32K | Details |
| Agentic codingwith OpenCode | Borderline | 4-bit | 10–13 tok/s | ≈3 min | 32K | Details |
| Long documents (64K)with Open WebUI | Acceptable | 4-bit | 9–12 tok/s | ≈8 min | 64K | Details |
Speeds are estimates calibrated against measured benchmarks on similar hardware. “First reply” is the time to read the workload’s opening prompt. How we calculate
Get it running
Copy-paste commands for the 4-bit build with a 64K context. Ollama and llama.cpp download the file straight from Hugging Face.
ollama run hf.co/unsloth/gemma-4-31B-it-GGUF:Q4_K_M
Downloads 18.32 GB (Q4_K_M). Ollama's default context is small: set OLLAMA_CONTEXT_LENGTH=65536 before `ollama serve` for long prompts.
Other models for MacBook Pro M3 Max 36 GB
The best-rated models for chat on this machine.
- Qwen3.6 35B-A3B80–110 tok/sExcellent
- Gemma 4 26B-A4B55–70 tok/sExcellent
- LFM2.5 24B-A2B110–150 tok/sExcellent
- Qwen3.5 35B-A3B80–110 tok/sExcellent
- gpt-oss-20b65–90 tok/sExcellent
- GLM-4.7-Flash (30B-A3B)75–100 tok/sExcellent
Cheapest machines for Gemma 4 31B
Hardware rated comfortable or better for chat, lowest price first.
- Desktop PC · Intel Arc Pro B60 24GB · 32 GB RAM$1,500 · 9–16 tok/sComfortable
- Ryzen AI Max+ 395 mini PC / Framework Desktop · 64 GB$1,600 · 10–14 tok/sComfortable
- Desktop PC · Intel Arc Pro B60 24GB · 64 GB RAM$1,628 · 9–16 tok/sComfortable
- Desktop PC · GeForce RTX 3090 · 32 GB RAM$1,700 · 30–40 tok/sExcellent
- Desktop PC · Radeon RX 7900 XTX · 32 GB RAM$1,800 · 30–45 tok/sExcellent
- Desktop PC · GeForce RTX 3090 · 64 GB RAM$1,828 · 30–40 tok/sExcellent
Same chip, other memory: 96 GB
Different tool, context size or workload?
These pages use typical settings. The full check models your exact setup.