Back to leaderboard

qwen3.5-35b-a3b-t48-c8k:latest

OLLAMA GGUF

qwen35moe · 36.0B · Q4_K_M

Not Rec.

Sep 6, 2026 · Intel Xeon® E5-4650 v2

qwen3.8-27b-mlx

LM-STUDIO MLX

qwen3_5 · 27B · 8bit

Marginal

Aug 17, 2026 · Apple M5 Max

Global Score
81 vs 59
Hardware Fit
63 vs 63
Quality Score
88 vs 58

Hardware

qwen3.5-35b-a3b-t4… qwen3.8-27b-mlx
MachineDell Inc. PowerEdge R820MacBook Pro
CPUIntel Xeon® E5-4650 v2Apple M5 Max
Cores8018
RAM256 GB64 GB
GPUMicrosoft Remote Display Adapter, NVIDIA Tesla P4, NVIDIA Tesla P4, Microsoft Basic Display AdapterApple M5 Max
OSMicrosoft Windows Server 2022 Datacenter 10.0.20348macOS 27.0
Archx64arm64
Power Modebalancedbalanced

Performance

qwen3.5-35b-a3b-t4… qwen3.8-27b-mlx
Tokens/sec17.218.3
First chunk845 ms4 ms
TTFT845 ms530 ms
Load time103.1 sN/A
Memory usage21.9 GB38.5 GB
Memory %9%60%

HW Fit Score Breakdown

qwen3.5-35b-a3b-t48-c8k…

Speed
16/50
TTFT
17/20
Memory
30/30

qwen3.8-27b-mlx

Speed
23/50
TTFT
20/20
Memory
20/30

Quality

qwen3.5-35b-a3b-t48-c8k…

Reasoning
17/20
Coding
18/20
Instruction
16/20
Structured
15/15
Math
12/15
Multilingual
10/10
Reasoning: Strong Coding: Strong Instruction Following: Strong Structured Output: Strong Math: Strong Multilingual: Strong

qwen3.8-27b-mlx

Reasoning
19/20
Coding
14/20
Instruction
6/20
Structured
1/15
Math
13/15
Multilingual
5/10
Reasoning: Strong Coding: Adequate Instruction Following: Weak Structured Output: Poor Math: Strong Multilingual: Weak

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