
What happened
Sakana AI said on October 9 that its Japan-tuned LLM, Sakana Namazu, has been adopted by Evidence Finder, the doctor-facing evidence search tool from medical-AI company Iris.
Why it matters
Iris's search-and-verification algorithms combined with Sakana Namazu now power the part that selects, compares and synthesizes papers into answers for doctors.
WHO IT HITSDoctors using Iris's Evidence Finder will get answers drafted by Sakana Namazu, and medical-AI product teams at similar evidence tools may look at Iris's model choice as a reference.
Summaries like this, in your inbox every morning.
The adoption follows an evaluation process rather than a simple procurement. Iris says the Evidence Finder evaluation model using Sakana Namazu recorded a 96.4% accuracy rate on February's 120th National Medical Licensing Examination, which the company describes as the top score among domestic foundation models designated by METI's GENIAC.
According to Sakana AI, that examination result, together with a qualitative evaluation of actual answers where practical usefulness was recognized, led to the decision to adopt Sakana Namazu for Evidence Finder.
The division of labor is explicit: Iris's search-and-verification algorithms select, compare and integrate relevant papers for a doctor's question, while Sakana Namazu generates the answer. Evidence Finder, provided by the medical-AI company Iris, shows citations and includes a "Verify" feature that automatically checks the existence of the papers the AI cites.
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 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.
Johns Hopkins astrophysicist Brice Ménard used Anthropic's Claude Science to build the first complete ultravio…

Google Cloud announced Gemini エージェント at its Gemini at Work 2026 event on October 8, US time

Anthropic added Claude Dashboards, which turns company data into live dashboards, and Claude Motion, which cre…

Valuence's "Generative AI User Trends Survey" (June 2024–May 2026) tracked six services including ChatGPT, Gem…

OpenAI began offering a ChatGPT feature that accepts uploaded audio files and can transcribe them, summarize t…

On a self-built 76-question Jev-format set, six trained 3B–9B open-weight models (Imajev-4B, Clef-Flash 9B, Je…
