Increasingly, I am also leveraging other models like Deepseek, Qwen, and Minimax for the evaluator agent in my /goal automoously loops.
Same here. Happy with Opus 4.8 (planning) and GPT-5.5 (execution).
Also, breaking steps into smaller ones for increasing quality is so underrated. This is why dynamic workflows are a bigger deal than most people think.
Users praise the addition of Deepseek Qwen and Minimax models to autonomous AI loops because it works really well.
Yes, it works really well.
1/ Claude Fable drains subscription quotas and is too expensive at API cost (our team has spent over $2k in a single day).
We've found that cheaper models + adversarial review loops achieve similar (sometimes better) results at significantly lower cost. 🧵
@omarsar0 Hmmm and I see this too ...
Increasingly, I am also leveraging other models like Deepseek, Qwen, and Minimax for the evaluator agent in my /goal automoously loops.
Same here. Happy with Opus 4.8 (planning) and GPT-5.5 (execution).
Also, breaking steps into smaller ones for increasing quality is so underrated. This is why dynamic workflows are a bigger deal than most people think.