Are MoEs Completely Overrated? Reddit User Slams Low Execution Intelligence
infieldmitt · reddit · 2026-08-08
A Reddit user sparked debate by heavily criticizing Mixture of Experts (MoE) models, using Qwen 3.6 35B A3B as an example. They argued that since each token is generated by only 3B parameters, the model exhibits shockingly poor execution in coding tasks despite having the right ideas. The user called MoEs the 'worst of both worlds' and proposed an 'Accumulation of Experts' approach to dynamically scale intelligence.
More from Models
- Opus 5.5 turns heads with surprisingly polished vertical video demo — FinanceYF5 · 2026-10-16
- Astra 6.1 may arrive before end of month, reads dev's 'we are cooking' tease — haider1 · 2026-10-16
- Open-source OCR model indra-ocr trends on Hugging Face with CPU support — datalab-to · 2026-10-16
- Xiaohongshu's Vela 35B turns anonymized user behavior into supervision, gains +11.5 on VelaBench — AdinaYakup · 2026-10-16
- KOL: No better model than Anthropic's Opus right now, its magic is back — kimmonismus · 2026-10-16
- Users blast Gemini in Chrome's 'draconian' usage caps: memory costs 2-4% per use — North-American · 2026-10-16