Manus AI: real or hype?

Early observations of Manus AI and comparison to other AI tools

TLDR: If we look at the Hype scale (10 being “very hyped & it’s fake marketing”, 1 being “this is real and already very useful”), I shifted from 7–9 toward 3–5 after using the early access version. Manus is a real contender that takes one step towards “doing stuff for you” from “just answering questions”.

But Manus’ pricing model surfaces inevitable issues of “AI accountability” and “contract dispute” that we need to address as agentic solutions make their ways into our lives and work.

After Anthropic revealed Computer Use in October 2024. Manus AI, built by a team in China, surprised the community in March 2025. Unlike Anthropic’s developer-first approach, Manus is taking a bold step by promising a General Agent that’s consumer-facing from Day 1. Many were skeptical about its capabilities. The demo seemed too good to be true.

I received early access today (April 2, 2025). I am excited to peek and understand how it compares to popular, state-of-the-art solutions in the market, specifically Google’s Gemini, Anthropic’s Claude, Perplexity, and OpenAI’s ChatGPT.

Is Manus a real deal or just hype (or even fake marketing as some claimed)?

Before we dive in, here is the launch video from a month ago (~March 2025) in case you missed it.

Here are a few things to keep in mind as the premise and scope of this review and discussion:

• Use everyday use cases and UX instead of edge cases
• Not a technical review of model, data, and systems
• No bias towards Chinese vendors (I am a Chinese working closely with US-based AI vendors and based in Toronto, Canada)
• Include more Manus AI screenshots in this post since other tools are generally assessable and assuming most people are familiar. You can find public links to chat history using all tools at the end of the post.