Preferred Networks (PFN) Open Models: Model List, Licenses and Articles

September 26, 2026September 27, 2026

About Preferred Networks (PFN)

Preferred Networks (PFN) is a Japanese AI company. On its pfnet organization on Hugging Face it publishes PLaMo, a series of language models trained from scratch in-house with strong Japanese ability. As a domestic company that handles large language models from pre-training onward, it is a particularly familiar publisher for Japanese users.

What It Releases

• PLaMo 3 NICT: pre-trained (Base) models developed by PFN in collaboration with the National Institute of Information and Communications Technology (NICT), in 2B, 8B and 31B. According to the model card, they are trained on English and Japanese data and use a hybrid of sliding-window attention and traditional attention layers. A new version of the 31B was released in June 2026.
• PLaMo 2.1 VL: 2B and 8B models that also read images (March 2026).
• A small model for finance: plamo-3-610m-fin-instruct from September 2026 is a Japanese model of about 0.9B specialized for the financial domain (we have a family page).
• Others include models trained from Qwen for medicine and finance (Preferred-MedLLM, qfin).

Licenses

Most PLaMo models use the “PLaMo Community License". Any use beyond personal or academic use counts as commercial, and commercial use requires both of the following:

1. Prior registration on PFN’s official registration page
2. Annual revenue or sales of the user (or its affiliated company) of 1 billion yen or less

Otherwise, you need a separate commercial license from PFN. When providing PLaMo-based models or products to third parties, you must state “Built with PLaMo" and include “PLaMo" in the names of models you create. Models trained on PLaMo outputs are also treated as derived models under the same terms. On Hugging Face, some models require accepting the license before download.

Preferred-MedRECT-32B and plamo-2-1b, on the other hand, are Apache-2.0, and Preferred-MedLLM-Qwen-72B uses the original Qwen license. The Qwen-based qfin models for finance are under the PLaMo Community License. This summary is not legal advice; read the license text before use.

Running Them Locally

• There are no official GGUF builds (as of 2026-09-26). For llama.cpp and Ollama you will need third-party conversions.
• You can choose a size. The 0.9B to 8B models are easy to handle on consumer GPUs. For which quantization fits which GPU, including the 31B, see the memory tables in our articles.

Sources: the pfnet organization; the PLaMo Community License Agreement shown on pfnet/plamo-2.1-8b-vl; model cards of pfnet/plamo-3-nict-8b-base and pfnet/plamo-3-610m-fin-instruct; the model list of the pfnet organization on Hugging Face (all as of 2026-09-26).

Our Coverage and Data

Local Model Watch has published 1 article(s) on Preferred Networks (PFN)’s own models, 0 on third-party fine-tunes and quantized builds based on them, and 0 from its official blog. The model list below is collected by code from Hugging Face every day. Part of our publisher index; model families link to their family pages.

At a Glance

Recent Models on Hugging Face

The publisher lists 30 models in total on Hugging Face. Below are up to 15 of the most recent generative models (text, image, video, audio and so on), the scope of this site, per organization. Classifiers, feature extractors, research components and repositories without a model card are omitted, and the publisher’s own quantized builds and helper variants are folded into the row of the original model. Parameter counts are left blank for repositories that pack weights into integer types, where Hugging Face’s count is not the real parameter count.

From the Hugging Face API. Last changed 2026-09-26 (JST). Collected by code.

Licenses of Its Own Models (as recorded in our articles)

The license can differ from model to model even within one publisher. This is not legal advice; check each model’s license text before use.

By Model Family

Articles on Its Own Models

Last updated 2026-09-26 (JST). The explanation at the top of this page was written with the help of AI from the primary sources it cites. The tables and lists under “Our Coverage and Data" are assembled by code from our article log and the Hugging Face API.