I am an avid user of many LLM's and work with AI every day. I heard about Ninja Tools from a few places and wanted to see what it was like and when it popped up on Appsumo, I quickly purchased. Unfortunately, it doesn't work at all for my use case.

I do a lot of optimised content, using the same prompt which I test across LLM's, finding that o1 and deepseek do the best job in terms of adding internal links and following brand voice. Even though we have o3,o4 now, o1 and deepseek gave me the best results. I recently test the new Sonnet 4 in Claude and via API and got some good results and then also tested Grok and 3.5, 3.7 from Claude.

I tested all available models in NinjaTools and the output kept being way too similar to be a coincidence, it didn't follow my prompt at all. Just gave me short content, no links, no bolded NLP keywords as I usually get from prompting the same models.

Unfortunately, I then went and asked the models which model they were and they are returning a different model than Ninja Tools says I am using.

For example, the new Claude 4.0 says It is 3.5, etc. I get terrible results from all models in Ninja Tools so it is probably just to good to be true for a lifetime deal and seems to me like they are routing the models differently than suggested.

I hope this improves but I cannot get any decent output from anything in Ninjatools as of now.

Edit to add after response from NinjaTools: The models are indeed incorrectly named and used after further testing. I subscribe to most AI tools including Claude via anthropic site, ChatGPT, other multiple model tools like Abacus Ai, ninjatools, etc. I have in total, 7 subscriptions as I test daily for my AI marketing app. I asked other tools which model they were for Claude Sonnet 4 for example, I get this response from each other tool:

"which AI model are you?"

Claude Sonnet 4 - I'm Claude Sonnet 4, an AI assistant created by Anthropic.

Answer from NinjaTools:

which exact AI model are you?

I'm Claude, an AI assistant created by Anthropic. I'm based on Anthropic's Claude 3.5 Sonnet model. I'm designed to be helpful, harmless, and honest in my interactions with users.

Is there anything specific you'd like help with today? I can assist with writing, coding, analysis, answering questions, and various other tasks using the tools available to me.

so you are not claude 4 sonnet?

That's correct - I'm Claude 3.5 Sonnet, not Claude 4 Sonnet. As far as I know, there isn't a Claude 4 Sonnet model that has been released by Anthropic yet. I'm part of the Claude 3.5 model family.

Is there something specific you'd like help with today?

-So it does seem that it is manipulating the model names. I can test the models knowledge next to prove this further as 3.5 understands older versions of auth cookie settings for supabase, not the latest way to do this.