baidu/ernie-4.5-21b-a3b
LM-STUDIO GGUFernie4_5-moe · 21B · Q4_K_M
Nov 1, 2026 · AMD RYZEN AI MAX+ 395
exaone-3.5-2.4b-instruct-mlx
LM-STUDIO MLXexaone · 2.4B · 8bit
Mar 4, 2026 · Apple M4
Global Score
86 vs 75
Hardware Fit
93 vs 100
Quality Score
83 vs 64
Hardware
baidu/ernie-4.5-21… exaone-3.5-2.4b-in…
MachineGMKtec EVO-X3MacBook Air
CPUAMD RYZEN AI MAX+ 395Apple M4
Cores3210
RAM120 GB32 GB
GPUAMD Radeon(TM) 8060S GraphicsApple M4
OSMicrosoft Windows 11 Pro 10.0.26300macOS 26.3
Archx64arm64
Power Modebalancedbalanced
Performance
baidu/ernie-4.5-21… exaone-3.5-2.4b-in…
Tokens/sec46.337.0
First chunk33 ms288 ms
TTFT383 ms288 ms
Load time49.7 sN/A
Memory usage32.5 GB2.4 GB
Memory %27%8%
HW Fit Score Breakdown
baidu/ernie-4.5-21b-a3b
Speed
43/50
TTFT
20/20
Memory
30/30
exaone-3.5-2.4b-instruc…
Speed
50/50
TTFT
20/20
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
exaone-3.5-2.4b-instruc…
Reasoning
9/20
Coding
13/20
Instruction
15/20
Structured
14/15
Math
5/15
Multilingual
8/10
Reasoning: Weak Coding: Adequate Instruction Following: Strong Structured Output: Strong Math: Weak 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