starcoder2:3b

starcoder2 · 3B · Q4_0

MacBook Air (Apple M4)

32 GB · macOS 26.3

Tested on March 2, 2026

Top 89% Compare

Global Score

48 /100

Marginal

Hardware Fit

100/100

Quality

13/100

Get this model

🦙

Ollama

ollama pull starcoder2:3b

View on Ollama Library

ollama.com/library/starcoder2

🤗

Find on HuggingFace

GGUF versions & quantizations

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: 49.4
Standard deviation Variation of speed across benchmark runs — lower means more consistent: ±18.5
Time to first token How fast the model starts responding — lower is better: 155 ms
Load time Time to load the model into memory before inference: 0.4 s
Memory usage RAM consumed during inference — percentage of total system memory: 3.0 GB (9%)
Total tokens Total number of tokens generated across all benchmark prompts: 520

Score breakdown

Speed

40/40

Time to first token

30/30

Memory

30/30

Quality

Reasoning

0/20

Coding

2/20

Instruction following

5/20

Structured output

2/15

Math

2/15

Multilingual

2/10

Category levels

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

Metadata

Spec version: 0.2.0
Runtime: Ollama 0.17.4
Model format: GGUF
Hardware profile: BALANCED
Result hash: 14c0938227f89f1b26f378a0feed5f056f2d65ec2f579b488d700416d67f7c6a

Interpretation

Hardware fit: 100/100. Overall suitability: MARGINAL (Global 48/100). Category profile: Reasoning: Poor, Coding: Poor, Instruction Following: Weak, Structured Output: Poor, Math: Poor, Multilingual: Poor. Warning: model produced very low accuracy on quality tasks — results may be unusable despite good hardware performance.

Warnings

Model produced very low accuracy on quality tasks — results may be unusable despite good hardware performance.
Token speed is unstable (stddev 18.5 tok/s, mean 49.4 tok/s) — may indicate thermal throttling or memory pressure.

Run yours now

npx metrillm@latest bench

Requires Node 20+ and Ollama running