Skip to content

DEV Community

Hard Numbers Series' Articles

Back to Pitambar Mahato's Series

Cover image for I Tested 4 LLM APIs for Coding Work — All 20 Runs Passed, But Speed Varies 10x

Pitambar Mahato

Sep 3

I Tested 4 LLM APIs for Coding Work — All 20 Runs Passed, But Speed Varies 10x

#llm #benchmark #claude #coding

7 min read

Cover image for I Ran 3 Open-Weight LLMs Head-to-Head on a 24GB Mac — One Was 3x Faster

Pitambar Mahato

Sep 4

I Ran 3 Open-Weight LLMs Head-to-Head on a 24GB Mac — One Was 3x Faster

#llm #benchmark #opensource #ai

5 min read

Cover image for I Tested Q4_K_M vs MXFP4 on the Same Laptop — The Supposedly-Faster New Format Lost

Pitambar Mahato

Sep 5

I Tested Q4_K_M vs MXFP4 on the Same Laptop — The Supposedly-Faster New Format Lost

#llm #benchmark #performance #ai

6 min read

Cover image for I Tested 5 Local LLM Sizes in n8n. The 1.5B Model Beat 14B on Tool-Calling.

Pitambar Mahato

Sep 6

I Tested 5 Local LLM Sizes in n8n. The 1.5B Model Beat 14B on Tool-Calling.

#ai #opensource #llm #devops

7 min read

Cover image for Can Local LLMs Actually Call Functions? I Tested 2 Models on 6 Real API Tasks

Pitambar Mahato

Sep 7

Can Local LLMs Actually Call Functions? I Tested 2 Models on 6 Real API Tasks

#llm #aiagents #opensource #toolcalling

6 min read