Back to leaderboard

llama3.3:70b-instruct-q4_K_M

OLLAMA GGUF

llama · 70.6B · Q4_K_M

Not Rec.

Aug 20, 2026 · Intel Xeon® E5-2637 v3

codestral:22b

OLLAMA GGUF

llama · 22.2B · Q4_0

Good

Jun 26, 2026 · Cortex-X925

Global Score
51 vs 67
Hardware Fit
31 vs 67
Quality Score
60 vs 67

Hardware

llama3.3:70b-instr… codestral:22b
MachineDell Inc. Precision Tower 7910NVIDIA NVIDIA_DGX_Spark
CPUIntel Xeon® E5-2637 v3Cortex-X925
Cores1620
RAM96 GB122 GB
GPUNVIDIA GeForce RTX 3090, NVIDIA GeForce RTX 3060Device 2e12
OSMicrosoft Windows 10 Pro 10.0.19045Ubuntu 24.04.4 LTS
Archx64arm64
Power Modebalancedperformance

Performance

llama3.3:70b-instr… codestral:22b
Tokens/sec1.717.2
First chunk5553 ms129 ms
TTFT5.6 s129 ms
Load time0.4 s11.6 s
Memory usage50.3 GB18.9 GB
Memory %53%16%

HW Fit Score Breakdown

llama3.3:70b-instruct-q…

Speed
2/50
TTFT
5/20
Memory
24/30

codestral:22b

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

Quality

llama3.3:70b-instruct-q…

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

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

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