gptoss · 20.9B · MXFP4

gptoss · 20.9 · MXFP4

Mac Pro (Intel Core™ i5-10400)

32 GB · macOS 26.5.1

Tested on July 9, 2026 · Submitted by engeldlgado

Top 21% Compare

Global Score

84 /100

Excellent

Hardware Fit

69/100

Quality

91/100

Get this model

🦙

Ollama

ollama pull gptoss · 20.9B · MXFP4

View on Ollama Library

ollama.com/library/gptoss · 20.9B · MXFP4

Get it in LM Studio

Search and download models directly from the app

🤗

Find on HuggingFace

GGUF versions & quantizations

Hardware

Machine: Mac Pro
CPU: Intel Core™ i5-10400
Cores: 12 total (12 perf)
Frequency: 2.9 GHz
RAM: 32 GB Empty
GPU: AMD Radeon RX 6700 XT
OS: macOS 26.5.1
Arch: x64
Power Mode: balanced

Performance

Tokens/sec Tokens generated per second — higher is better: 17.5
Standard deviation Variation of speed across benchmark runs — lower means more consistent: ±4.7
First chunk latency Delay before the first streamed chunk arrives from runtime: 643 ms
Time to first token How fast the model starts responding — lower is better: 7.7 s
Load time Time to load the model into memory before inference: 0.0 s
Memory usage RAM consumed during inference — percentage of total system memory: 0.8 GB (3%)
Total tokens Total number of tokens generated across all benchmark prompts: 1886
Thinking tokens (est.) Estimated internal reasoning tokens — models with "thinking" generate more tokens: ~438

Score breakdown

Speed

34/50

Time to first token

5/20

Memory

30/30

Quality

Reasoning

19/20

Coding

19/20

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 ToshLLM (llama.cpp modified)
Model format: GGUF
Hardware profile: BALANCED
Result hash: 0de6f1666e83c662f6a62a9617d9b39253e02d513d05cb5cfbde1f5e87ae68d7

Interpretation

Hardware fit: 69/100. Overall suitability: EXCELLENT (Global 84/100). Category profile: Reasoning: Strong, Coding: Strong, Instruction Following: Adequate, Structured Output: Strong, Math: Strong, Multilingual: Strong.

Bench Environment

Thermal: nominal Power: AC CPU load: avg 25% (peak 31%)

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