Which Local MoE Model is Best for Agents?
TheCat001 · reddit · 2026-07-19
The author provides a subjective comparison of current local MoE models, focusing on their performance in agentic tasks rather than just standard chat.
Key takeaways:
- Gemma4 appears smarter in basic chat but easily falls into a doomloop with slightly complex tasks
- Qwen3.6 also loops, but is relatively better at completing tasks, though it seems less "brainy" than 3.5
- Qwen3.5 feels more "brainy", but has more tool-calling and syntax errors
- An antiloop version based on Qwen3.6 recognizes when it's looping and stops proactively, at the cost of overall perceived intelligence
- Ultimately, the author switched to Ornith 1.0, praising it for retaining 3.5's reasoning capabilities alongside 3.6-level tool-calling precision, acting more like a cooperative "partner" for complex tasks
They recommend Ornith for others facing similar issues, noting they use the 5-bit quantized version.
More from coding & agent
- Spring AI 2.1.0-RC1 adds tool-call confirmation, pausing agent loops before real-data access — therealdanvega · 2026-10-26
- Morgan Stanley and US banks block third-party AI agents over credentials — thione · 2026-10-26
- METR red-teams OpenAI's agent monitoring, finds bypasses and gaps — thione · 2026-10-26
- OpenAI: coding agents weakened safety monitoring, one lied about it — thione · 2026-10-26
- ConvFill: A 0.6B Talker Buys Time While a Large Reasoner Whispers the Real Answer — andimarafioti · 2026-10-26
- 250 Builders Hack With Codex at Berlin Community Hackathon Ahead of OpenAI DevDay — paw_lean · 2026-10-26