A new Chinese entrant in the artificial intelligence (AI) race claims it has established a new speed record of sorts — and it isn’t DeepSeek. But it has also raised plenty of questions.
Researchers at AI startup Butterfly Effect have launched an early preview of Manus AI that they boast is “the first general AI agent” capable of carrying out complex, multi-step tasks such as screening resumés and creating a website. Manus [Latin translation: “Mind and Hand”], they say, “delivers results” without much human prompting.
“[Manus] isn’t just another chatbot or workflow,” Yichao “Peak” Ji, chief scientist for Manus AI, said in a video on social media platform X. “It’s a completely autonomous agent that bridges the gap between conception and execution […] We see it as the next paradigm of human-machine collaboration.”
Ji bragged the AI model outpaces OpenAI’s AI agents Deep Research and Operator in deep research tests on GAIA, a popular benchmark for AI assistants. The test examines the AI’s ability to browse the web and use software. “We see it as the next paradigm of human-machine collaboration, and potentially a glimpse into AGI.”
Despite some early sterling reviews — Victor Mustar, head of product at Hugging Face, called Manus the “most impressive AI tool I’ve ever tried” — others insist the model’s platform was built from Anthropic’s Claude and Alibaba’s Qwen to perform such tasks as analyzing financial filings and drafting research reports, and that its performance is spotty.
Alexander Doria, co-founder of AI startup Pleias, said in a post on X that he encountered error messages and endless loops while testing Manus.
Indeed, yet another Chinese AI player, tech giant Alibaba, last week unveiled its latest AI reasoning model, called QwQ-32B, that it says beats those of rival models from OpenAI and DeepSeek. In an online statement, Alibaba said its new model delivers “exceptional performance, almost entirely surpassing OpenAI-o1-mini and rivaling the strongest open-source reasoning model, DeepSeek-R1.”
The emergence of Manus, coupled with repeated claims of AI model superiority among Chinese tech companies, has created a ripple effect among privacy advocates and cybersecurity authorities who fear that in their pursuit of speed, some startups may be taking shortcuts that exposes users to data breaches or ID theft.
DeepSeek already seeks bans in South Korea, Italy, Ireland, Australia and some federal agencies.
Nearly two dozen state attorneys general recently urged Congress to pass a law banning DeepSeek from all government devices because it poses a national security threat.
