hf.co/Edge-Quant/Nanbeige4.1-3B-Q8_0-GGUF:Q8_0

OLLAMA GGUF

llama · 3.93B · unknown

Not Rec.

Aug 22, 2026 · Apple M1 Pro

mistralai/devstral-small-2-2512

LM-STUDIO MLX

mistral3 · 24B · 4bit

Good

Mar 4, 2026 · Apple M4

Global Score

52 vs 69

Hardware Fit

75 vs 50

Quality Score

42 vs 77

Hardware

hf.co/Edge-Quant/N… mistralai/devstral…

MachineMacBook ProMacBook Air

CPUApple M1 ProApple M4

Cores1010

RAM16 GB32 GB

GPUApple M1 ProApple M4

OSmacOS 26.6.2macOS 26.3

Archarm64arm64

Power Modebalancedbalanced

Performance

hf.co/Edge-Quant/N… mistralai/devstral…

Tokens/sec43.55.7

First chunk225 ms5007 ms

TTFT30.0 s5.0 s

Load time2.7 sN/A

Memory usage8.2 GB13.2 GB

Memory %51%41%

HW Fit Score Breakdown

hf.co/Edge-Quant/Nanbei…

Speed

50/50

TTFT

0/20

Memory

25/30

mistralai/devstral-smal…

Speed

11/50

TTFT

9/20

Memory

30/30

Quality

hf.co/Edge-Quant/Nanbei…

Reasoning

13/20

Coding

1/20

Instruction

7/20

Structured

1/15

Math

13/15

Multilingual

7/10

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

mistralai/devstral-smal…

Reasoning

16/20

Coding

18/20

Instruction

13/20

Structured

14/15

Math

6/15

Multilingual

10/10

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

View hf.co/Edge-Quant/N… details | View mistralai/devstral… 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