Best Local LLMs for 12GB RAM

Ranked open-weight models that run well on a 12GB machine, with estimated speed and the exact ollama command for each.

Try: Qwen3 8B, RTX 4090, MacBook Pro M4, 16GB, iPhone 17 Pro

Best local models for ~12GB

12 picks

Estimates assume a representative Apple-Silicon machine with 12GB unified memory. Tok/s are ModelFit estimates, not measured benchmarks. Run the wizard for figures tuned to your exact chip.

Qwen3.5 9B Instruct

Qwen / 9B / Q4_K_M / ~7 GB

Best for: Quality, Coding, Reasoning·Perf: ~22.4 tok/s · first token ~0.9s

Runs wellC · 17

This model may feel memory-heavy on 12 GB RAM, but it is still listed for balanced speed and quality.

ollama

$ollama run qwen3.5:9b

Qwen3 8B

Qwen / 8B / Q4_K_M / ~6.5 GB

Best for: Chat, Coding·Perf: ~25.2 tok/s · first token ~0.8s

Runs wellC · 23

This model may feel memory-heavy on 12 GB RAM, but it is still listed for balanced speed and quality.

ollama

$ollama run qwen3:8b-q4_K_M

LFM2.5 8B-A1B

LFM2 / 8.3B / Q4_K_M / ~5.5 GB

Best for: On-device agents, tool calling, multilingual chat·Perf: ~57.1 tok/s · first token ~0.6s

Runs wellB · 35

Best for on-device agents, tool calling, multilingual chat. Strong fit for 12 GB RAM with balanced speed and quality.

ollama

$ollama run lfm2.5:8b-a1b-q4_K_M

Ornith 1.0 9B

Ornith / 9B / Q4_K_M / ~5.6 GB

Best for: Agentic coding on small machines·Perf: ~22.4 tok/s · first token ~0.9s

Runs wellB · 33

Best for agentic coding on small machines. Strong fit for 12 GB RAM with balanced speed and quality.

ollama

$ollama run ornith:9b

MiniCPM-V 4.5 8B

MiniCPM / 8.7B / Q4_K_M / ~5.7 GB

Best for: Vision, Multimodal·Perf: ~23.2 tok/s · first token ~0.9s

Runs wellB · 32

Best for vision, multimodal. Strong fit for 12 GB RAM with balanced speed and quality.

ollama

$ollama run minicpm-v4.5:8b

Llama 3.1 8B Instruct

Llama / 8B / Q4_K_M / ~6.5 GB

Best for: Chat, Coding·Perf: ~25.2 tok/s · first token ~0.8s

Runs wellC · 23

This model may feel memory-heavy on 12 GB RAM, but it is still listed for balanced speed and quality.

ollama

$ollama run llama3.1:8b-instruct-q4_K_M

Qwen2.5 Coder 7B

Qwen / 7B / Q4_K_M / ~5.5 GB

Best for: Coding·Perf: ~28.8 tok/s · first token ~0.8s

Runs wellB · 35

Best for coding. Strong fit for 12 GB RAM with balanced speed and quality.

ollama

$ollama run qwen2.5-coder:7b

DeepSeek-R1 Distill Qwen 7B

DeepSeek / 7B / Q4_K_M / ~5.5 GB

Best for: Reasoning, Coding·Perf: ~28.8 tok/s · first token ~0.8s

Runs wellB · 35

Best for reasoning, coding. Strong fit for 12 GB RAM with balanced speed and quality.

ollama

$ollama run deepseek-r1:7b

Qwen3.5 4B Instruct (Q8)

Qwen / 4B / Q8_0 / ~4.3 GB

Best for: Coding, Agents, Multimodal·Perf: ~27.6 tok/s · first token ~0.8s

Runs wellB · 49

Best for coding, agents, multimodal. Strong fit for 12 GB RAM with balanced speed and quality.

ollama

$ollama run qwen3.5:4b-q8_0

Granite 4.2 8B

Granite / 8.8B / Q4_K_M / ~5 GB

Best for: Chat, Coding·Perf: ~22.9 tok/s · first token ~0.9s

Runs wellB · 40

Best for chat, coding. Strong fit for 12 GB RAM with balanced speed and quality.

ollama

$ollama run granite4.2:8b

Qwen2.5 7B Instruct

Qwen / 7B / Q4_K_M / ~5.5 GB

Best for: Chat, Coding·Perf: ~28.8 tok/s · first token ~0.8s

Runs wellB · 35

Best for chat, coding. Strong fit for 12 GB RAM with balanced speed and quality.

ollama

$ollama run qwen2.5:7b-instruct-q4_K_M

Mistral 7B Instruct

Mistral / 7B / Q4_K_M / ~5.5 GB

Best for: Chat, Coding·Perf: ~28.8 tok/s · first token ~0.8s

Runs wellB · 35

Best for chat, coding. Strong fit for 12 GB RAM with balanced speed and quality.

ollama

$ollama run mistral:7b-instruct-q4_K_M

Browse by RAM tier

8 GB 16 GB 24 GB 32 GB 48 GB 64 GB 96 GB 128 GB

Want an exact recommendation?

The wizard tunes picks and speed estimates to your exact device, chip, and RAM.

Open ModelFit Wizard Browse all devices