qwen3.5-35b-a3b-uncensored-hauhaucs-aggressive

LM-STUDIO GGUF

qwen35moe · 35B · Q4_K_M

Good

Apr 21, 2026 · Apple M2 Max

falcon3-1b-instruct

LM-STUDIO MLX

llama · 1B · 3bit

Marginal

Mar 4, 2026 · Apple M4

Global Score

71 vs 54

Hardware Fit

92 vs 100

Quality Score

62 vs 34

Hardware

qwen3.5-35b-a3b-un… falcon3-1b-instruct

MachineMac StudioMacBook Air

CPUApple M2 MaxApple M4

Cores1210

RAM96 GB32 GB

GPUApple M2 MaxApple M4

OSmacOS 26.1macOS 26.3

Archarm64arm64

Power Modebalancedbalanced

Performance

qwen3.5-35b-a3b-un… falcon3-1b-instruct

Tokens/sec59.8121.8

First chunk12 ms176 ms

TTFT2.4 s176 ms

Load timeN/AN/A

Memory usage29.7 GB0.7 GB

Memory %31%2%

HW Fit Score Breakdown

qwen3.5-35b-a3b-uncenso…

Speed

50/50

TTFT

12/20

Memory

30/30

falcon3-1b-instruct

Speed

50/50

TTFT

20/20

Memory

30/30

Quality

qwen3.5-35b-a3b-uncenso…

Reasoning

18/20

Coding

8/20

Instruction

13/20

Structured

6/15

Math

12/15

Multilingual

5/10

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

falcon3-1b-instruct

Reasoning

7/20

Coding

1/20

Instruction

12/20

Structured

7/15

Math

2/15

Multilingual

5/10

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

View qwen3.5-35b-a3b-un… details | View falcon3-1b-instruct details

Run yours and compare

$ npm install -g metrillm@latest

$ metrillm

Requires Node 20+ and Ollama or LM Studio running

Or run without installing: npx metrillm@latest