Back to leaderboard

codestral:22b

OLLAMA GGUF

llama · 22.2B · Q4_0

Good

Jun 26, 2026 · Cortex-X925

qwen3.5:4b

OLLAMA GGUF

qwen35 · 4.7B · Q4_K_M

Marginal

Mar 5, 2026 · Apple M4

Global Score
67 vs 48
Hardware Fit
67 vs 77
Quality Score
67 vs 35

Hardware

codestral:22b qwen3.5:4b
MachineNVIDIA NVIDIA_DGX_SparkMac mini
CPUCortex-X925Apple M4
Cores2010
RAM122 GB16 GB
GPUDevice 2e12Apple M4
OSUbuntu 24.04.4 LTSmacOS 26.3
Archarm64arm64
Power Modeperformancebalanced

Performance

codestral:22b qwen3.5:4b
Tokens/sec17.218.9
First chunk129 ms316 ms
TTFT129 ms13.0 s
Load time11.6 s3.7 s
Memory usage18.9 GB5.9 GB
Memory %16%37%

HW Fit Score Breakdown

codestral:22b

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

qwen3.5:4b

Speed
43/50
TTFT
4/20
Memory
30/30

Quality

codestral:22b

Reasoning
11/20
Coding
17/20
Instruction
9/20
Structured
14/15
Math
6/15
Multilingual
10/10
Reasoning: Adequate Coding: Strong Instruction Following: Weak Structured Output: Strong Math: Weak Multilingual: Strong

qwen3.5:4b

Reasoning
10/20
Coding
1/20
Instruction
4/20
Structured
3/15
Math
8/15
Multilingual
9/10
Reasoning: Weak Coding: Poor Instruction Following: Poor Structured Output: Poor Math: Adequate Multilingual: Strong
View codestral:22b details | View qwen3.5:4b 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