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

THINKING MODEL

MacBook Air (Apple M4)

32 GB · macOS 26.3

Tested on March 5, 2026

Top 78% Compare

Global Score

53 /100

Marginal

Hardware Fit

99/100

Quality

33/100

Get this model

Get it in LM Studio

Search and download models directly from the app

🤗

Find on HuggingFace

MLX versions & conversions

Hardware

Machine: MacBook Air
CPU: Apple M4
Cores: 10 total (4 perf + 6 eff)
Frequency: 2.4 GHz
RAM: 32 GB LPDDR5
GPU: Apple M4
OS: macOS 26.3
Arch: arm64
Power Mode: balanced

Performance

Tokens/sec Tokens generated per second — higher is better: 25.4
Standard deviation Variation of speed across benchmark runs — lower means more consistent: ±0.2
First chunk latency Delay before the first streamed chunk arrives from runtime: 266 ms
Time to first token How fast the model starts responding — lower is better: 266 ms
Load time Time to load the model into memory before inference: N/A
Memory usage RAM consumed during inference — percentage of total system memory: 0.0 GB (0%)
Total tokens Total number of tokens generated across all benchmark prompts: 1530

Score breakdown

Speed

49/50

Time to first token

20/20

Memory

30/30

Quality

Reasoning

10/20

Coding

7/20

Instruction following

1/20

Structured output

1/15

Math

11/15

Multilingual

3/10

Category levels

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

Metadata

Spec version: 0.2.1
Runtime: LM Studio 0.4.6+1
Model format: GGUF
Hardware profile: BALANCED
Result hash: 602208c0e3e08998b36de8024396d24d8e7b61b4d0fee75956d842440f611934

Interpretation

Hardware fit: 99/100. Overall suitability: MARGINAL (Global 53/100). Category profile: Reasoning: Adequate, Coding: Weak, Instruction Following: Poor, Structured Output: Poor, Math: Adequate, Multilingual: Weak.

Bench Environment

Power: AC CPU load: avg 13% (peak 14%)

Run yours now

$ npm install -g metrillm@latest

$ metrillm

Requires Node 20+ and Ollama or LM Studio running

Or run without installing: npx metrillm@latest