GPT-20B:20b

gptoss · 20.9B · MXFP4

THINKING MODEL

HP HyperX OMEN Transcend Gaming Laptop 14-ga0xxx5CD612DXML (NVIDIA RTX Spark N1X (6144-core GPU, 20-core ) RTX Spark N1X (6144-core GPU, 20-core ))

55 GB · Microsoft Windows 11 Pro 10.0.28000

Tested on August 22, 2026

Top 2% Compare

Global Score

93 /100

Excellent

Hardware Fit

91/100

Quality

94/100

Get this model

🦙

Ollama

ollama pull GPT-20B:20b

View on Ollama Library

ollama.com/library/GPT-20B

Get it in LM Studio

Search and download models directly from the app

🤗

Find on HuggingFace

GGUF versions & quantizations

Hardware

Machine: HP HyperX OMEN Transcend Gaming Laptop 14-ga0xxx5CD612DXML
CPU: NVIDIA RTX Spark N1X (6144-core GPU, 20-core ) RTX Spark N1X (6144-core GPU, 20-core )
Cores: 20 total (20 perf)
Frequency: 4 GHz
RAM: 55 GB LPDDR5
GPU: NVIDIA RTX Spark N1X (6144-core Blackwell RTX GPU), NVIDIA NPU
OS: Microsoft Windows 11 Pro 10.0.28000
Arch: arm64
Power Mode: balanced

Performance

Tokens/sec Tokens generated per second — higher is better: 57.8
Standard deviation Variation of speed across benchmark runs — lower means more consistent: ±1.0
First chunk latency Delay before the first streamed chunk arrives from runtime: 203 ms
Time to first token How fast the model starts responding — lower is better: 3.0 s
Load time Time to load the model into memory before inference: 19.2 s
Memory usage RAM consumed during inference — percentage of total system memory: 12.0 GB (22%)
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: ~670

Score breakdown

Speed

50/50

Time to first token

11/20

Memory

30/30

Quality

Reasoning

19/20

Coding

17/20

Instruction following

18/20

Structured output

15/15

Math

15/15

Multilingual

10/10

Category levels

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

Metadata

Spec version: 0.2.1
Runtime: Ollama 0.32.15
Model format: GGUF
Hardware profile: HIGH-END
Result hash: 0580df79f7f68230bc09b2b67ca6559ea7b8f3b16974fd948a638bc291520fb4

Interpretation

Hardware fit: 91/100. Overall suitability: EXCELLENT (Global 93/100). Category profile: Reasoning: Strong, Coding: Strong, Instruction Following: Strong, Structured Output: Strong, Math: Strong, Multilingual: Strong.

Warnings

CPU appears throttled (2.9 GHz current vs 4.0 GHz nominal, 73%).

Bench Environment

Power: AC CPU load: avg 20% (peak 26%)

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