mlx-community/Nanbeige4.1-3B-8bit

LM-STUDIO GGUF

Marginal

Mar 5, 2026 · Apple M4

hf.co/unsloth/Qwen3.8-27B-GGUF:UD-IQ1_M

OLLAMA GGUF

qwen35 · 27.3B · unknown

Good

Aug 21, 2026 · AMD RYZEN AI MAX+ 395

Global Score

53 vs 68

Hardware Fit

99 vs 69

Quality Score

33 vs 67

Hardware

mlx-community/Nanb… hf.co/unsloth/Qwen…

MachineMacBook AirAZW GTR Pro

CPUApple M4AMD RYZEN AI MAX+ 395

Cores1032

RAM32 GB125 GB

GPUApple M4Radeon 8060S

OSmacOS 26.3Ubuntu 24.04.4 LTS

Archarm64x64

Power Modebalancedlow-power

Performance

mlx-community/Nanb… hf.co/unsloth/Qwen…

Tokens/sec25.421.1

First chunk266 ms571 ms

TTFT266 ms571 ms

Load timeN/A4.9 s

Memory usage0.0 GB22.7 GB

Memory %0%18%

HW Fit Score Breakdown

mlx-community/Nanbeige4…

Speed

49/50

TTFT

20/20

Memory

30/30

hf.co/unsloth/Qwen3.8-2…

Speed

20/50

TTFT

19/20

Memory

30/30

Quality

mlx-community/Nanbeige4…

Reasoning

10/20

Coding

7/20

Instruction

1/20

Structured

1/15

Math

11/15

Multilingual

3/10

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

hf.co/unsloth/Qwen3.8-2…

Reasoning

12/20

Coding

10/20

Instruction

10/20

Structured

15/15

Math

10/15

Multilingual

10/10

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

View mlx-community/Nanb… details | View hf.co/unsloth/Qwen… 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