meta/muse-glimmer

muse-glimmer

THINKING MODEL

LENOVO 82XV (Intel Gen Intel® Core™ i7-13620H)

32 GB · Microsoft Windows 11 Pro 10.0.26200

Tested on August 17, 2026 · Submitted by riche

Top 100% Compare

Global Score

9 /100

Not Rec.

Hardware Fit

28/100

Quality

1/100

Get this model

Get it in LM Studio

Search and download models directly from the app

🤗

Find on HuggingFace

MLX versions & conversions

Hardware

Machine: LENOVO 82XV
CPU: Intel Gen Intel® Core™ i7-13620H
Cores: 16 total (16 perf)
Frequency: 2.4 GHz
RAM: 32 GB DDR5
GPU: Intel(R) UHD Graphics, NVIDIA GeForce RTX 4060 Laptop GPU
OS: Microsoft Windows 11 Pro 10.0.26200
Arch: x64
Power Mode: balanced

Performance

Tokens/sec Tokens generated per second — higher is better: 3.2
Standard deviation Variation of speed across benchmark runs — lower means more consistent: ±0.0
First chunk latency Delay before the first streamed chunk arrives from runtime: 19 ms
Time to first token How fast the model starts responding — lower is better: 68.2 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: 16.9 GB (53%)
Total tokens Total number of tokens generated across all benchmark prompts: 1658
Thinking tokens (est.) Estimated internal reasoning tokens — models with "thinking" generate more tokens: ~857

Score breakdown

Speed

5/50

Time to first token

0/20

Memory

23/30

Quality

Reasoning

0/20

Coding

0/20

Instruction following

0/20

Structured output

0/15

Math

1/15

Multilingual

0/10

Category levels

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

Metadata

Spec version: 0.2.1
Runtime: LM Studio
Model format: GGUF
Hardware profile: BALANCED
Result hash: fe07524e4fc4ec779c27e3943ff7534650ad7616b7e41b6e35b86b71a6b2b530

Interpretation

Hardware fit: 28/100. Overall suitability: NOT RECOMMENDED (Global 9/100). Category profile: Reasoning: Poor, Coding: Poor, Instruction Following: Poor, 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.
Model memory footprint is estimated via LM Studio CLI rather than measured from a fresh load.

Disqualifiers

Token speed too low: 3.2 tok/s (minimum: 5 tok/s for BALANCED profile)
Time to first token too high: 68170ms (maximum: 18480ms for BALANCED profile)

Bench Environment

Power: AC CPU load: avg 40% (peak 47%)

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