baidu/ernie-4.5-21b-a3b
LM-STUDIO GGUFernie4_5-moe · 21B · Q4_K_M
Nov 1, 2026 · AMD RYZEN AI MAX+ 395
qwen3.5-35b-a3b-t48-c8k:latest
OLLAMA GGUFqwen35moe · 36.0B · Q4_K_M
Sep 6, 2026 · Intel Xeon® E5-4650 v2
Global Score
86 vs 81
Hardware Fit
93 vs 63
Quality Score
83 vs 88
Hardware
baidu/ernie-4.5-21… qwen3.5-35b-a3b-t4…
MachineGMKtec EVO-X3Dell Inc. PowerEdge R820
CPUAMD RYZEN AI MAX+ 395Intel Xeon® E5-4650 v2
Cores3280
RAM120 GB256 GB
GPUAMD Radeon(TM) 8060S GraphicsMicrosoft Remote Display Adapter, NVIDIA Tesla P4, NVIDIA Tesla P4, Microsoft Basic Display Adapter
OSMicrosoft Windows 11 Pro 10.0.26300Microsoft Windows Server 2022 Datacenter 10.0.20348
Archx64x64
Power Modebalancedbalanced
Performance
baidu/ernie-4.5-21… qwen3.5-35b-a3b-t4…
Tokens/sec46.317.2
First chunk33 ms845 ms
TTFT383 ms845 ms
Load time49.7 s103.1 s
Memory usage32.5 GB21.9 GB
Memory %27%9%
HW Fit Score Breakdown
baidu/ernie-4.5-21b-a3b
Speed
43/50
TTFT
20/20
Memory
30/30
qwen3.5-35b-a3b-t48-c8k…
Speed
16/50
TTFT
17/20
Memory
30/30
Quality
baidu/ernie-4.5-21b-a3b
Reasoning
16/20
Coding
17/20
Instruction
15/20
Structured
14/15
Math
12/15
Multilingual
9/10
Reasoning: Strong Coding: Strong Instruction Following: Strong Structured Output: Strong Math: Strong Multilingual: Strong
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
Run yours and compare
$
npm install -g metrillm@latest$
metrillmRequires Node 20+ and Ollama or LM Studio running
Or run without installing: npx metrillm@latest