Qwen falls apart at tool-calling while gemma4 shines in a DIY agentic loop

spammmmmmmmy · reddit · 2026-09-23

A Reddit user shares hands-on experience building a simple DIY agentic loop: the model is told its available tools, proposes commands with a RUN: prefix, and the user approves or edits them each turn.

Key findings:

They conclude there may be a trade-off slider between "creative" and "obedient" behavior, and open a discussion on which models handle tool-calling best.

Original post →

More from coding & agent

coding & agent channel →