openai/gpt-oss-20b

gpt_oss · 20B · MXFP4

THINKING MODEL

MacBook Pro (Apple M5 Max)

64 GB · macOS 27.0

Tested on July 17, 2026 · Submitted by itxjobe

Top 25% Compare

Global Score

83 /100

Not Rec.

Hardware Fit

80/100

Quality

84/100

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 Pro
CPU: Apple M5 Max
Cores: 18 threads (6 cores)
Frequency: 2.4 GHz
RAM: 64 GB LPDDR5
GPU: Apple M5 Max
OS: macOS 27.0
Arch: arm64
Power Mode: balanced

Performance

Tokens/sec Tokens generated per second — higher is better: 131.7
Standard deviation Variation of speed across benchmark runs — lower means more consistent: ±1.3
First chunk latency Delay before the first streamed chunk arrives from runtime: 7 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: N/A
Memory usage RAM consumed during inference — percentage of total system memory: 15.8 GB (25%)
Total tokens Total number of tokens generated across all benchmark prompts: 1708
Thinking tokens (est.) Estimated internal reasoning tokens — models with "thinking" generate more tokens: ~942

Score breakdown

Speed

50/50

Time to first token

0/20

Memory

30/30

Quality

Reasoning

18/20

Coding

13/20

Instruction following

14/20

Structured output

14/15

Math

15/15

Multilingual

10/10

Category levels

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

Metadata

Spec version: 0.2.1
Runtime: LM Studio
Model format: MLX
Hardware profile: HIGH-END
Result hash: 45a48f4ef95f465edf7b66fe4eccc4b559905ccf487cab316933fbfb07f793d3

Interpretation

Hardware fit: 80/100. Overall suitability: NOT RECOMMENDED (Global 83/100). Category profile: Reasoning: Strong, Coding: Adequate, Instruction Following: Adequate, Structured Output: Strong, Math: Strong, Multilingual: Strong.

Warnings

Model memory footprint is estimated via LM Studio CLI rather than measured from a fresh load.

Disqualifiers

Time to first token too high: 30000ms (maximum: 14977ms for HIGH-END profile)

Bench Environment

Power: AC CPU load: avg 32% (peak 35%)

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