Names without a link are companies not yet registered in OpenBizGraph
§05
Network
NETWORK
Loading the relationship graph...
Drag to move · click to open
§06
History
25 EVENTS
Since its founding in 2014, the company has pursued the mission of 'making the real world computable.'
Partnered with the Motivation Switch Group to launch the Playgram™ programming curriculum.
Released Playgram Typing, a typing practice curriculum for children.
Group company Preferred Computational Chemistry (now Matlantis Inc.) began offering the general-purpose atom-level simulator Matlantis™.
Launched PFN 3D Scan, a high-quality 3D scanning service for businesses.
Released HAPiiBOT™, a compact floor-cleaning robot jointly developed by group company Preferred Robotics and Amano.
Preferred Robotics released the autonomous transport robot Kachaka™.
Successfully automated the butadiene extraction plant at ENEOS Kawasaki Refinery using an AI system.
In August 2023, ranked 3rd worldwide (out of 954 teams) in a Kaggle competition that identified contrails in satellite imagery to mitigate climate change impacts.
Released the open-source Japanese large language model PLaMo™-13B.
In October 2023, achieved 4th and 5th place worldwide (1st and 2nd in Japan) out of 2,662 teams in the Kaggle competition 'LLM Science Exam' evaluating the accuracy of science quiz answers generated by LLMs.
In November 2023, secured 5th and 9th place worldwide (out of 616 teams) in the Kaggle competition 'Fast or Slow? Predict AI Model Runtime' predicting AI model execution speed.
In 2024, celebrated the company's 10th anniversary.
Launched P-FEP, a computational service that predicts binding activity values for low-molecular-weight compounds in drug discovery research.
Initiated AI-driven automation of the crude oil processing plant at ENEOS Kawasaki Refinery.
Started offering PLaMo Lite, a lightweight language model for edge devices.
Announced PreferredAI™, a product and service utilizing generative AI.
Launched MiseMise™, an operational improvement solution for chain stores.
Launched the AI cloud service Preferred Computing Platform™ (PFCP™) powered by the AI processor MN-Core 2.
In October 2024, ranked 10th worldwide (out of 1,915 teams) in the Kaggle competition 'RSNA 2024 Lumbar Spine Degenerative Classification' that classifies degenerative lumbar spine conditions using MR images.
On October 21, 2024, hosted a 10th-anniversary event in Tokyo to showcase PFN's achievements over the past decade and future initiatives.
Introduced the flagship model PLaMo Prime of PLaMo™.
Entered into a capital partnership with SBI Group, Sekisui House, Japan Bank for International Cooperation, Mitsubishi Corporation, and Wacom.
Formed capital alliances with Kodansha, Sekisui House, TBS, Toei Animation, Sumitomo Mitsui Trust Bank, Mitsubishi UFJ Trust Bank, and others.
Established a capital partnership with Shin-Etsu Chemical and Sumitomo Life Insurance.
§07
News
40 ARTICLES
PFN's Mouse Tail Vein Automatic Injection Robot AUTiv Wins the 12th Robot Award "Chairman's Award" of the Japan Machinery Industry Federation
The mouse tail vein automatic injection system AUTiv, developed by Preferred Networks Inc., received the "Chairman's Award" at the 12th Robot Award. It was praised for its high-precision drug delivery, reduced puncture errors, and lowered burden on researchers and laboratory animals, and has been delivered to about 20 units domestically. AUTiv is the world’s first robot system that combines deep learning and precise positioning control to immobilize mice without anesthesia.
Domestic Finance-Specialized Small Language Model PLaMo 3.0 Fin Open-Weights Release
Preferred Networks released the small language model PLaMo 3.0 Fin, designed for financial institutions, with open weights. The model is available in cloud isolated environments and on-premises, and post-training uses the company’s own AI semiconductor MN-Core 2. In the demo, it classified the urgency of customer inquiries and outputted results in JSON format.
Preferred Networks’ AI Writing Guidance System "Ronjirou" Wins e-Learning Award
The AI writing guidance system "Ronjirou", developed by Preferred Networks (PFN), received the AI Special Award (Learning Support) at the 23rd Japan e‑Learning Awards. The system utilizes the domestic foundation model PLaMo™ to support middle and high school students in writing essays and creative writing, and AI provides logical connections and improvement suggestions within one minute, reducing the instructor’s grading burden.
PFN Releases PLaMo 3 Translate LLM
Preferred Networks, Inc. (PFN) announced the open‑weight release of PLaMo 3 Translate, a suite of translation language models available in four sizes ranging from 0.8 B to 31 B parameters. The models can translate locally on smartphones or internal servers, bypassing the cloud, and are free to download and use under the PLaMo Community License for individuals and enterprises with annual revenue below 1 billion Japanese yen.
PFN Releases Open-Weight AI Translation Model PLaMo 3 Translate for Edge Use
Preferred Networks (PFN) has released the AI translation model PLaMo 3 Translate as an open-weight model that can be used on edge devices without relying on the cloud. The model is available in four sizes ranging from 0.8B to 31B, and companies or individuals with revenue or sales of 1 billion yen or less in the most recent fiscal year can use it free of charge under the PLaMo Community License.
PFN to Showcase MN-Core Series at SC26
Preferred Networks, Inc. (PFN) will exhibit at the SC26 International Supercomputing Conference in Chicago, Illinois, from November 15 to 20, 2026. The company’s booth (4606) will display physical models and mock-ups of the MN‑Core™ L1000 AI inference chip under development, outline the roadmap for the MN‑Core L Series, and provide live demonstrations using the MN‑Core 2 Devkit.
PFN Exhibits MN-Core Series at SC26
Preferred Networks, Inc. (PFN) will exhibit at the international conference SC26, held in Chicago, Illinois, USA, from November 15 to 20, 2026. At booth 4606, they will showcase a structural model and mock-up of the inference-oriented AI semiconductor MN-Core™︎ L1000, future plans, a live demo using the MN-Core 2 Devkit, and research and development of a highly efficient AI computing platform.
PFN Opens Applications for 2028 Graduates
Preferred Networks has started its second call for applications for full-time positions for 2028 graduates. The company seeks students graduating in 2028 who are driven to address real-world challenges, and describes work spanning diverse projects that use advanced computer science expertise and AI tools such as coding agents.
PFN Opens Applications for New Graduate Positions Starting in 2028
Preferred Networks (PFN) has opened the first round of applications for new graduate positions for people expected to graduate in 2028. Positions include engineers, researchers, and materials discovery roles. The application deadline for the first round is November 30, 2026, at 11:59 p.m., and offers are expected to be made in December.
ZenRing Group and PFN enter business capital partnership to fuse geospatial information and AI
ZenRing Group and Preferred Networks (PFN) have entered a business capital partnership to fuse high-precision geospatial information with AI that has a time axis, aiming to create services that grasp, predict, and support decision-making in the real world. As part of the partnership, ZenLin and ZenLin DataCom will underwrite PFN's third-party allotment of capital, promoting mid- to long-term collaboration from planning and development to commercialization. The background is that the evolution of generative AI and multimodal AI has made it possible to handle location information and temporal changes cross-sectionally, but high-quality data and specialized expertise are deemed essential for solving real-world problems.
