qwen/qwen3.6-35b-a3b

qwen35moe · 35B · Q4_K_M

Micro-Star International Co., Ltd. MS-7E51 (AMD Ryzen 7 9700X 8-Core Processor)

31 GB · Microsoft Windows 11 Pro 10.0.26200

Tested on August 19, 2026

Top 23% Compare

Global Score

85 /100

Excellent

Hardware Fit

74/100

Quality

89/100

Get this model

Get it in LM Studio

Search and download models directly from the app

🤗

Find on HuggingFace

MLX versions & conversions

Hardware

Machine: Micro-Star International Co., Ltd. MS-7E51
CPU: AMD Ryzen 7 9700X 8-Core Processor
Cores: 16 total (16 perf)
Frequency: 3.8 GHz
RAM: 31 GB DDR5
GPU: NVIDIA GeForce RTX 5080, AMD Radeon(TM) Graphics
OS: Microsoft Windows 11 Pro 10.0.26200
Arch: x64
Power Mode: balanced

Performance

Tokens/sec Tokens generated per second — higher is better: 22.1
Standard deviation Variation of speed across benchmark runs — lower means more consistent: ±28.5
First chunk latency Delay before the first streamed chunk arrives from runtime: 9 ms
Time to first token How fast the model starts responding — lower is better: 588 ms
Load time Time to load the model into memory before inference: N/A
Memory usage RAM consumed during inference — percentage of total system memory: 20.5 GB (66%)
Total tokens Total number of tokens generated across all benchmark prompts: 1231

Score breakdown

Speed

37/50

Time to first token

20/20

Memory

17/30

Quality

Reasoning

16/20

Coding

18/20

Instruction following

18/20

Structured output

15/15

Math

12/15

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: LM Studio 0.4.15+2
Model format: GGUF
Hardware profile: BALANCED
Result hash: f3e6491ac72a6b2f32cda4280de50499727f48c5ac5469a4e67e3b1b849d3e59

Interpretation

Hardware fit: 74/100. Overall suitability: EXCELLENT (Global 85/100). Category profile: Reasoning: Strong, Coding: Strong, Instruction Following: Strong, Structured Output: Strong, Math: Strong, Multilingual: Strong.

Warnings

Token speed is unstable (stddev 28.5 tok/s, mean 22.1 tok/s) — may indicate thermal throttling or memory pressure.
Model memory footprint is estimated via LM Studio CLI rather than measured from a fresh load.

Bench Environment

CPU load: avg 40% (peak 48%)

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