HudHud Tech — Tech & AI news platform
Monday, 16 November 2026
Breaking
Trends | Artificial Intelligence

Moonshot AI Targets Scientific Discovery with Kimi Models

16 November 2026 3 min read هدهد تك | Hudhud Tech

Chinese artificial intelligence start-up Moonshot AI, the developer of the Kimi models, is focusing on scientific discovery, joining US rivals in the effort to transform AI into a research engine.

The Beijing-based company is increasingly exploring applications that require advanced AI reasoning capabilities, particularly in biology, chemistry, and pharmaceutical research.

Giacomo Ficari, Moonshot’s director of international, stated during a keynote address at the Web Summit technology conference in Lisbon that “the real excitement is … science.” He noted that while finance and legal sectors are at a similar level of intelligence as coding, science is pushing the boundaries even further.

Ficari explained that coding has driven much of the recent demand for advanced AI models and is nearing saturation in terms of technological progress. This has prompted Moonshot to investigate applications in scientific and pharmaceutical research.

He highlighted that while OpenAI is focusing heavily on mathematics, Moonshot is concentrating on biology and chemistry, emphasizing the potential advancements in pharmaceuticals.

Moonshot has begun to implement its ambitions. Recently, DP Technology, a Beijing-based AI-for-science start-up, announced a collaboration with Moonshot to integrate the Kimi K3.1 Pro model into its research agent for biomedical applications.

DP Technology reported that in an internal evaluation of AI models for biological research tasks, the Kimi K3.1 Pro combined with its coding tool achieved the highest score compared to other undisclosed model combinations.

In one demonstration, DP Technology’s research agent, BioMaster, and Kimi analyzed gene-expression data from 3,696 mouse pancreatic cells, identifying 100 candidate genes for further investigation in approximately 90 minutes, significantly faster than the estimated four hours of human work.

While these results showcase how advanced models can automate time-consuming data analysis, they do not demonstrate the ability to make autonomous scientific breakthroughs.

This distinction was emphasized in a recent paper co-authored by Moonshot founder Yang Zhilin and Turing Award-winning computer scientist Andrew Yao. The study examined whether leading AI models could contribute to original mathematical research.

The researchers introduced a benchmark consisting of 100 tasks, including mathematical theorem proving and attempts to solve longstanding unsolved problems in theoretical computer science.

They found that while frontier models could produce verifiable results and assist with complex research tasks, none could independently solve the benchmark’s open mathematical conjectures, highlighting both the potential and current limitations of AI in autonomous research.

Moonshot did not immediately respond to a request for comment.

The company’s push into scientific research comes amid rival claims of scientific milestones from US peers. Last month, OpenAI announced in a blog post that its models had made three new discoveries across mathematics, computer science, and physics, generating proofs for long-standing problems. However, these results faced criticism from the scientific community, particularly among mathematicians.

In September, Anthropic claimed its researchers had used its Claude model to discover a new enzyme system.

Despite these advancements, industry insiders remain cautious about AI’s near-term impact on pharmaceutical research and development. A survey conducted by Citi in June among 50 US-based pharmaceutical R&D executives found that only 24 percent expected AI to significantly improve the probability of drug success, even though 72 percent indicated their firms were scaling or had fully scaled the technology across research operations.

 

Source: SCMP Tech (Asia)

What do you think of this story?

Share this story: