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

llama · 3.93B · unknown

THINKING MODEL

MacBook Pro (Apple M1 Pro)

16 GB · macOS 26.6.2

Tested on August 22, 2026

Top 87% Compare

Global Score

52 /100

Not Rec.

Hardware Fit

75/100

Quality

42/100

Get this model

🦙

Ollama

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

View on Ollama Library

ollama.com/library/hf.co/Edge-Quant/Nanbeige4.1-3B-Q8_0-GGUF

Get it in LM Studio

Search and download models directly from the app

🤗

Find on HuggingFace

GGUF versions & quantizations

Hardware

Machine: MacBook Pro
CPU: Apple M1 Pro
Cores: 10 total (8 perf + 2 eff)
Frequency: 2.4 GHz
RAM: 16 GB LPDDR5
GPU: Apple M1 Pro
OS: macOS 26.6.2
Arch: arm64
Power Mode: balanced

Performance

Tokens/sec Tokens generated per second — higher is better: 43.5
Standard deviation Variation of speed across benchmark runs — lower means more consistent: ±0.0
First chunk latency Delay before the first streamed chunk arrives from runtime: 225 ms
Time to first token How fast the model starts responding — lower is better: 30.0 s
Load time Time to load the model into memory before inference: 2.7 s
Memory usage RAM consumed during inference — percentage of total system memory: 8.2 GB (51%)
Total tokens Total number of tokens generated across all benchmark prompts: 1445
Thinking tokens (est.) Estimated internal reasoning tokens — models with "thinking" generate more tokens: ~930

Score breakdown

Speed

50/50

Time to first token

0/20

Memory

25/30

Quality

Reasoning

13/20

Coding

1/20

Instruction following

7/20

Structured output

1/15

Math

13/15

Multilingual

7/10

Category levels

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

Metadata

Spec version: 0.2.1
Runtime: Ollama 0.32.14
Model format: GGUF
Hardware profile: ENTRY
Result hash: 656b5498c571c2b7e7b26bb376a6dc2c339967c92c02d9bcac353d51e6f74a74

Interpretation

Hardware fit: 75/100. Overall suitability: NOT RECOMMENDED (Global 52/100). Category profile: Reasoning: Adequate, Coding: Poor, Instruction Following: Weak, Structured Output: Poor, Math: Strong, Multilingual: Adequate.

Disqualifiers

Time to first token too high: 30000ms (maximum: 22068ms for ENTRY profile)

Bench Environment

Power: AC CPU load: avg 37% (peak 38%)

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