Best Local LLMs for 48GB RAM

Ranked open-weight models that run well on a 48GB machine — with estimated speed and the exact ollama command for each.

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

Best local models for ~48GB

12 picks

Estimates assume a representative Apple-Silicon machine with 48GB unified memory. Tok/s are ModelFit estimates, not measured benchmarks. Run the wizard for figures tuned to your exact chip.

01

Qwen3.6 35B-A3B

Qwen / 35B / Q4_K_M / ~22 GB

Best for: Reasoning, Coding, Agents·Perf: ~42.7 tok/s (est.) · first token ~1.5s

Runs well

Best for reasoning, coding, agents. Strong fit for 48 GB RAM with balanced speed and quality.

ollama
$ollama run qwen3.6:35b-a3b
02

Qwen3.5 35B-A3B Instruct

Qwen / 35B / Q4_K_M / ~20 GB

Best for: Reasoning, Coding, Agent scenarios·Perf: ~42.7 tok/s (est.) · first token ~1.5s

Runs well

Best for reasoning, coding, agent scenarios. Strong fit for 48 GB RAM with balanced speed and quality.

ollama
$ollama run qwen3.5:35b-a3b
03

Qwen3.5 27B Instruct

Qwen / 27B / Q4_K_M / ~16 GB

Best for: Chat, Coding, Complex reasoning·Perf: ~53.9 tok/s (est.) · first token ~0.6s

Runs well

Best for chat, coding, complex reasoning. Strong fit for 48 GB RAM with balanced speed and quality.

ollama
$ollama run qwen3.5:27b
04

Qwen3.6 27B

Qwen / 27B / Q4_K_M / ~18 GB

Best for: Coding, Quality, Long context·Perf: ~53.9 tok/s (est.) · first token ~0.6s

Runs well

Best for coding, quality, long context. Strong fit for 48 GB RAM with balanced speed and quality.

ollama
$ollama run qwen3.6:27b
05

Gemma 4 26B-A4B

Gemma / 26B / Q4_K_M / ~16 GB

Best for: Chat, Coding, Multimodal·Perf: ~55.8 tok/s (est.) · first token ~0.6s

Runs well

Best for chat, coding, multimodal. Strong fit for 48 GB RAM with balanced speed and quality.

ollama
$ollama run gemma4:26b
06

LFM2 24B-A2B Instruct

LFM2 / 24B / Q4_K_M / ~14 GB

Best for: Local AI agents, privacy-first tool calling, MCP workflows·Perf: ~59.9 tok/s (est.) · first token ~0.6s

Runs well

Best for local ai agents, privacy-first tool calling, mcp workflows. Strong fit for 48 GB RAM with balanced speed and quality.

ollama
$ollama run lfm2:24b-a2b
07

Qwen3 14B

Qwen / 14B / Q4_K_M / ~11 GB

Best for: Coding, Quality·Perf: ~97.4 tok/s (est.) · first token ~0.6s

Perfect fit

Best for coding, quality. Strong fit for 48 GB RAM with balanced speed and quality.

ollama
$ollama run qwen3:14b-q4_K_M
08

Gemma 4 31B

Gemma / 31B / Q4_K_M / ~20 GB

Best for: Quality, Coding, Multimodal·Perf: ~47.6 tok/s (est.) · first token ~1.5s

Runs well

Best for quality, coding, multimodal. Strong fit for 48 GB RAM with balanced speed and quality.

ollama
$ollama run gemma4:31b
09

Qwen2.5 Coder 14B

Qwen / 14B / Q4_K_M / ~11 GB

Best for: Coding·Perf: ~97.4 tok/s (est.) · first token ~0.6s

Perfect fit

Best for coding. Strong fit for 48 GB RAM with balanced speed and quality.

ollama
$ollama run qwen2.5-coder:14b
10

Qwen3 30B

Qwen / 30B / Q4_K_M / ~22 GB

Best for: Quality, Coding·Perf: ~49 tok/s (est.) · first token ~1.5s

Runs well

Best for quality, coding. Strong fit for 48 GB RAM with balanced speed and quality.

ollama
$ollama run qwen3:30b
11

Qwen2.5 14B Instruct

Qwen / 14B / Q4_K_M / ~11 GB

Best for: Coding, Chat·Perf: ~97.4 tok/s (est.) · first token ~0.6s

Perfect fit

Best for coding, chat. Strong fit for 48 GB RAM with balanced speed and quality.

ollama
$ollama run qwen2.5:14b-instruct-q4_K_M
12

DeepSeek-R1 Distill Qwen 14B

DeepSeek / 14B / Q4_K_M / ~11 GB

Best for: Reasoning, Quality·Perf: ~97.4 tok/s (est.) · first token ~0.6s

Perfect fit

Best for reasoning, quality. Strong fit for 48 GB RAM with balanced speed and quality.

ollama
$ollama run deepseek-r1:14b

Browse by RAM tier

8 GB16 GB24 GB32 GB48 GB64 GB96 GB128 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