AIToday
Large Language ModelsAI Business & IndustryTHE DECODERPublished: Nov 17, 2026, 06:01 JST

Moonshot AI shifts Kimi models from coding to science, Ficari says

Send on LINE
Moonshot AI shifts Kimi models from coding to science, Ficari says

3 Key Points

  1. What happened

    Moonshot AI is redirecting its Kimi language models from coding toward biology, chemistry, and pharma research; Giacomo Ficari said at Web Summit in Lisbon that science demands smarter AI than law, finance, or code.

  2. Why it matters

    The pivot points Kimi at work that, by Moonshot's own framing, asks more of a model than coding does — a harder target than the coding benchmarks it is stepping away from.

WHO IT HITSThe move lands on research scientists and R&D teams in biology, chemistry, and pharma who would use AI assistants like BioMaster for tasks such as gene analysis. It may also affect enterprise software teams that had been evaluating Kimi for coding work, though the body does not say whether coding support continues.

Not sure about something? Ask the AI

Questions and answers are published on this page.

Summaries like this, every morning.

Context & Analysis

The pivot comes as Moonshot AI sees diminishing returns in coding, where benchmark gains have flattened. Rather than keep chasing those scores, the company is repositioning Kimi toward domains it argues are harder for AI: biology, chemistry, and pharma research. That claim rests on the idea that scientific work demands smarter models than law, finance, or code — a framing Giacomo Ficari offered at Web Summit in Lisbon.

Early adoption is already taking shape. Beijing startup DP Technology is integrating Kimi K3.1 Pro into BioMaster, its research assistant, and a test on gene data from 3,696 mouse pancreatic cells suggests speed gains over manual human analysis. Yet the body is clear that fully independent breakthroughs remain distant: on the TCSExam benchmark, leading models produced verifiable proofs but could not solve any open conjectures alone. US rivals face similar limits — Anthropic's Claude found an enzyme system in September, while mathematicians criticized OpenAI's claimed math and physics breakthroughs.

The commercial audience is not fully convinced. A June Citi survey of 50 US research leaders found 72 percent already using AI broadly, but only 24 percent expect it to noticeably improve drug development success rates. That gap between adoption and expected impact is the backdrop against which Moonshot's science-first bet will be judged.

FAQ
What efficiency did Kimi K3.1 Pro show in the BioMaster test?
In one test, the AI analyzed gene data from 3,696 mouse pancreatic cells and flagged 100 candidate genes in 90 minutes, versus an estimated four hours of human work.
Can AI models make fully independent scientific breakthroughs yet?
No. On TCSExam, a 100-problem theoretical CS benchmark, leading models produced verifiable proofs but could not crack any open conjectures alone.
Is the pharma industry convinced AI will improve drug development?
Not yet. A June Citi survey of 50 US research leaders found 72 percent already using AI broadly, but only 24 percent expect it to noticeably improve drug development success rates.
Send on LINE

AI news that matters for your work, delivered every morning.

Pick your industry and the AI tools you use, and get news related to your work every day.

Free · 30 seconds with Google · unsubscribe anytimeWhat is AIToday? →

Ask AI

Ask AI anything about this article. The AI reads this article, earlier AIToday articles, and Wikipedia, and cites its sources. Q&As are published on this page for other readers too.

Questions and answers are published on this page.

Related Articles

Next articleIntel, CEO and CFO sued after 28.6 percent stock plunge