Iqidis.AI is better than Grok 3 and Claude 3.7 "Thinking": Here's Why

From Abhivardhan, our President

Okay, so I was a bit bored, which is why I tried Researcher by Iqidis.AI.

This tool, like Claude 3.7 "Thinking", Grok 3, and "Deep Research" by Perplexity, offers three important deliverables - Quick Question, Standard Research, and Deep Research.

Here's my review.

1️⃣ Yes, this is a rather better tool as opposed to "Deep Research" by Perplexity and kind of equitably comparable to Grok 3. Is the subscription rate worth it? Can't comment on that. Why? Well, Researcher by Iqidis is accessible for FREE use.

.

2️⃣ What did I like: first off, it uses the "chain of thought" approach which of course has its limitations as proven in the case of enough CS/AI literature. However, unlike Grok or Perplexity, the tool somehow remain specific in searching academic sources when it requires to and web sources, when it needs to. Like Andriy Burkov said about Perplexity, I think a lot of times both Deep Research, and Claude 3.7 "Thinking" create an illusion of reading or "gobbling up" too many sources.

Researcher by Iqidis at least doesn't waste its chain-of-thought process by gobbling too many sources. It refers 30-50 sources, but then they are assorted quite better, or at least even if we say chain-of-thought by design is illusory - the tool attempts at bringing some sense in chain-of-thought by virtue of each steps.

3️⃣ The specificity in source choicing between academic and web sources makes it time-saving, a bit more reliable than Grok and Perplexity, and even Claude, interestingly.

4️⃣ Both tabular and points-based analysis that it offers doesn't seem like hallucinating, directly. The text design shows it scrapes specifically and gives output in the most specific sense, ever. Yeah, it might mention case laws and provisions more specifically - but it still feels workable to me.

Overall, Devansh Devansh has developed quite a nuanced workflow around the same chain-of-thought architecture, so I'd recommend this tool over Perplexity Deep Research, OpenAI's useless GPT 4.1 and Grok 3 (because while Grok gobbles up a lot, its tenacity to be too much all-comprehensive, even if it is more precise in collating information via pattern matching than even Claude - makes it inconsistent, and suffers from hallucination).

Do try researcher.iqidis.ai and use it at your wisdom. I've given some screenshots, so open to discuss about it. Maybe Devansh might also like doing that.