LM Studio
a friendly desktop chat window
Runs on Mac, Windows, Linux. The lights below are for Gemma 4 31B, the strongest model this app runs on our list.
How good is it?
Close to DeepSeek (July 2026)
A local chat model can be wonderful at everyday writing and still miss today's very best cloud model on hard reasoning.
Will it fit?
A green light means it fits and feels fine. Amber means it fits, with a wait or a tight memory pool. Red means the file is bigger than the room we count.
Small Mac
Won't fitMac mini M4 16 GB
It does not fit
Roomy Mac
Runs slowMac mini M4 32 GB
Slow. You will wait on each reply
Around $1,099 · estimate
Source: Apple M4 Mac mini memory options, 14 Oct 2026
Big Mac
Runs wellMac mini M5 Pro 64 GB
Usable, with a short wait
Around $2,499 · estimate
Source: Apple Mac mini specs, 14 Oct 2026
Small graphics card
Won't fitRTX 4060 8 GB
It does not fit
Mid graphics card
Won't fitRTX 5070 Ti 16 GB
It does not fit
Big graphics card
Tight fitRTX 4090 24 GB
Faster than you can read
Around $2,900 all-in · estimate
Source: NVIDIA RTX 4090 specifications, 14 Oct 2026
Mini PC
Runs wellStrix Halo mini PC 64 GB
Usable, with a short wait
Around $2,199 · estimate
Source: Compute Market roundup, 14 Oct 2026
Official install guides
We link to the makers. We do not write our own install steps.
Other apps that run Gemma 4 31B
- Ollama's download page · Ollama · Mac, Windows, Linux
- LM Studio's app guide · LM Studio · Mac, Windows, Linux
- AnythingLLM's docs · AnythingLLM · Mac, Windows, Linux
Models this app runs
- Gemma 4 31B
- Qwen3.8 27B
- Gemma 4 26B A4B
- Qwen3 235B-A22B
- Qwen3.5 27B
- Qwen3.5 35B-A3B
- Gemma 3 27B
- Qwen3-Coder 30B-A3B
- gpt-oss 120B
- Gemma 3 12B
- DeepSeek R1 Distill 32B
- Qwen3 8B
- Gemma 3n E4B
- Llama 3.3 70B
- gpt-oss 20B
- Gemma 3 4B
- Qwen2.5-Coder 32B
- Aya Expanse 32B
- Phi-4 14B
- Qwen2.5-Coder 7B
- Aya Expanse 8B
- Llama 3.2 3B
- olmOCR 2 7B
- Qwen2.5-VL 7B