gpt-oss:20b

gptoss · 20.9B · MXFP4

THINKING MODEL

Mac mini (Apple M4)

24 GB · macOS 26.3

Tested on March 10, 2026

Top 11% Compare

Global Score

83 /100

Excellent

Hardware Fit

65/100

Quality

90/100

Get this model

🦙

Ollama

ollama pull gpt-oss:20b

View on Ollama Library

ollama.com/library/gpt-oss

Get it in LM Studio

Search and download models directly from the app

🤗

Find on HuggingFace

GGUF versions & quantizations

Hardware

Machine: Mac mini
CPU: Apple M4
Cores: 10 total (4 perf + 6 eff)
Frequency: 2.4 GHz
RAM: 24 GB LPDDR5
GPU: Apple M4
OS: macOS 26.3
Arch: arm64
Power Mode: balanced

Performance

Tokens/sec Tokens generated per second — higher is better: 17.1
Standard deviation Variation of speed across benchmark runs — lower means more consistent: ±2.7
First chunk latency Delay before the first streamed chunk arrives from runtime: 705 ms
Time to first token How fast the model starts responding — lower is better: 10.2 s
Load time Time to load the model into memory before inference: 23.1 s
Memory usage RAM consumed during inference — percentage of total system memory: 13.1 GB (55%)
Total tokens Total number of tokens generated across all benchmark prompts: 1688
Thinking tokens (est.) Estimated internal reasoning tokens — models with "thinking" generate more tokens: ~823

Score breakdown

Speed

37/50

Time to first token

5/20

Memory

23/30

Quality

Reasoning

19/20

Coding

18/20

Instruction following

13/20

Structured output

15/15

Math

15/15

Multilingual

10/10

Category levels

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

Metadata

Spec version: 0.2.1
Runtime: Ollama 0.17.7
Model format: GGUF
Hardware profile: ENTRY
Result hash: 825daaba450a8a7e74dd94eff4dfee29e6b2fb8675d89d23cca6ad0971a4dfe0

Interpretation

Hardware fit: 65/100. Overall suitability: EXCELLENT (Global 83/100). Category profile: Reasoning: Strong, Coding: Strong, Instruction Following: Adequate, Structured Output: Strong, Math: Strong, Multilingual: Strong.

Warnings

Significant swap activity during benchmark (+9.6 GB). Model may exceed available RAM — results are severely degraded.

Bench Environment

Power: AC Swap delta: +9.6 GB CPU load: avg 15% (peak 20%)

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