Comparison of Clay's Argon vs. Sonnet 3.7 Model Output Issues

Does Clay think Argon is a better model than Sonnet 3.7 (it costs the same) and when I accidentally run it, I get outputs like this. Just seems weird Clay's internal "best" recommended model would do this. https://app.clay.com/workspaces/9621/workbooks/wb_WzQNJm94zKfp/tables/t_WmAeiNUwvwc5/views/gv_kMhv3zZvQib6

Channeled
APP
·
·
We recommend Claygent Argon as our strongest overall model for deep research and complex analysis. It significantly outperforms Opus and GPT 4o. However, each model has its strengths, and the best choice depends on your specific task. Argon costs 3 credits per row, like our other Claygent models (Neon and Helium). If you're experiencing unexpected outputs, it could be due to various factors. To help you better, could you please share more details about the specific task and the output you received? This will allow us to provide more targeted assistance.
Source: Claygent: AI Web Scraper [Clay 101] - Clay University
Daniela D.
·
·
Hey Jordan! Thanks for flagging. Happy to help. We'll review those fields and get back with an update🙏🏾. Let us know if you have additional examples of any weird outputs returned by Argon.
ishrak h.
·
·
Yeah Helium is smarter, faster, and so much better than Argon i have no idea whats wrong with Argon
Daniela D.
·
·
Hey Ishrak! Thanks for the message. Could you share an example of weird results you've gotten from Argon? Happy to review this too.