qwen/qwen3.8-27b
LM-STUDIO MLXqwen3_5 · 27B · 8bit
Aug 18, 2026 · Apple M4 Max
devstral-small-2-24b-instruct-2512
LM-STUDIO GGUFmistral3 · 24B · IQ4_XS
Sep 20, 2026 · AMD Ryzen 7 5800X 8-Core Processor
Global Score
81 vs 81
Hardware Fit
68 vs 80
Quality Score
87 vs 81
Hardware
qwen/qwen3.8-27b devstral-small-2-2…
MachineMacBook ProSystem manufacturer
CPUApple M4 MaxAMD Ryzen 7 5800X 8-Core Processor
Cores1616
RAM128 GB31 GB
GPUApple M4 MaxNavi 44 [Radeon RX 9060 XT]
OSmacOS 26.5.2Ubuntu 26.04.1 LTS
Archarm64x64
Power Modebalancedperformance
Performance
qwen/qwen3.8-27b devstral-small-2-2…
Tokens/sec16.421.5
First chunk5 ms12 ms
TTFT742 ms212 ms
Load timeN/AN/A
Memory usage38.5 GB15.9 GB
Memory %30%51%
HW Fit Score Breakdown
qwen/qwen3.8-27b
Speed
18/50
TTFT
20/20
Memory
30/30
devstral-small-2-24b-in…
Speed
35/50
TTFT
20/20
Memory
25/30
Quality
qwen/qwen3.8-27b
Reasoning
17/20
Coding
18/20
Instruction
16/20
Structured
15/15
Math
11/15
Multilingual
10/10
Reasoning: Strong Coding: Strong Instruction Following: Strong Structured Output: Strong Math: Adequate Multilingual: Strong
devstral-small-2-24b-in…
Reasoning
16/20
Coding
17/20
Instruction
17/20
Structured
14/15
Math
7/15
Multilingual
10/10
Reasoning: Strong Coding: Strong 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