mistral:latest

OLLAMA GGUF

llama · 7.2B · Q4_K_M

Good

Mar 2, 2026 · Apple M4

qwen3.5-4b-mlx

LM-STUDIO MLX

qwen3_5 · 4B · 4bit

Good

Mar 3, 2026 · Apple M4

Global Score

73 vs 71

Hardware Fit

96 vs 88

Quality Score

63 vs 63

Hardware

mistral:latest qwen3.5-4b-mlx

MachineMacBook AirMacBook Air

CPUApple M4Apple M4

Cores1010

RAM32 GB32 GB

GPUApple M4Apple M4

OSmacOS 26.3macOS 26.3

Archarm64arm64

Power Modebalancedbalanced

Performance

mistral:latest qwen3.5-4b-mlx

Tokens/sec22.523.1

First chunkN/A3279 ms

TTFT257 ms3.3 s

Load time0.6 sN/A

Memory usage10.1 GB2.9 GB

Memory %32%9%

HW Fit Score Breakdown

mistral:latest

Speed

36/50

TTFT

30/20

Memory

30/30

qwen3.5-4b-mlx

Speed

45/50

TTFT

13/20

Memory

30/30

Quality

mistral:latest

Reasoning

8/20

Coding

12/20

Instruction

15/20

Structured

14/15

Math

5/15

Multilingual

9/10

Reasoning: Weak Coding: Adequate Instruction Following: Strong Structured Output: Strong Math: Weak Multilingual: Strong

qwen3.5-4b-mlx

Reasoning

15/20

Coding

16/20

Instruction

6/20

Structured

6/15

Math

11/15

Multilingual

9/10

Reasoning: Adequate Coding: Strong Instruction Following: Weak Structured Output: Weak Math: Adequate Multilingual: Strong

View mistral:latest details | View qwen3.5-4b-mlx details

Run yours and compare

$ npm install -g metrillm@latest

$ metrillm

Requires Node 20+ and Ollama or LM Studio running

Or run without installing: npx metrillm@latest