Q: Why did you lower quality of the LLMs?
I purchased your highest tier but unfortunately must return , based on that decision. Users asked questions about the low amount of credits provided in the deal and was something that I completely agreed with them on, especially with no credit rollover. But your answer was to remove the pro language model (keeping only deepseek)? These flash models are not performing the same way as the pro models. If anything, you should have added the flash models and kept the prompt models alongside it. This move to present models that don't perform as well and are therefore cheaper in response to not providing enough credits is a bit too concerning for me to continue with the product. It feels like anytime you need to lower cost you will simply lower quality to compensate (rather on appsumo or otherwise).
Arjun_RetrieverAI
Aug 11, 2026A: Hey TetanicRaptor, we removed Gemini Pro because its an outdated model and it was just way too costly inference wise. The Gemini Flash Fast option is the latest Gemini model released so far.
We just got accepted into AWS, Azure, Nvidia, GMI Cloud, Thinking Machines startup programs and so will have a lot more model options offered soon.
We found the DeepSeek Flash mode to perform pretty well and its usually the prompting itself not being clear that is the issue.
We significantly updated the cloud platform and I will check if there was a side effect of this. Our browser extension is usually the most reliable path.
As we've learned in this business, new and better cannot be confused. The new gemini model is not performing as well as the old pro model. While I'm glad the platform has joined startup programs, you removed models, to save on usage cost in the middle of an active campaign, which changes it's performance. It's an easy way to spook buyers as it indicates financial hardships & instability
I understand cost and as an early adopter I expect things to break from time to time in this stage. I just need to be able to justify the purchase and it takes spending some time with the product without a variable such as which LLM will exist next week. I still the product. But this has been a bit of a rug slip.