PLaMo (Preferred Networks)
PFN's from-scratch Japanese LLM, now in Japan's government AI stack
Overview
PLaMo is Preferred Networks' domestically-built foundation model family — PLaMo Prime (flagship API), PLaMo Lite (edge), and PLaMo Translate, which Japan's Digital Agency deployed inside its government AI environment "Gennai" in December 2025 and expanded through 2026.
Pricing
Pricing shown for reference only. These figures reflect RECATOOLS research as of 12 Jul 2026 and may be out of date or incomplete. This is not financial or purchasing advice — always confirm the current price on the provider’s official website before making any decision.
- Web-based access
- Japanese-language optimized
- ¥1,000/1M output tokens
- ¥1,000 signup credit (launch campaign)
- Drop-in for openai-python / LangChain
- Private deployment options
- Finance-tuned and edge variants
What you can produce with PLaMo (Preferred Networks)
- PLaMo Prime — flagship commercial API model (OpenAI-API-compatible)
- PLaMo Lite — lightweight model for edge deployment
- PLaMo Translate — Japanese-specialized translation model, deployed in government use
- Finance-tuned PLaMo variant
- PLaMo Chat — free consumer chat assistant
- Open-weight models on Hugging Face (1B–100B, PLaMo 3 NICT series)
- vLLM inference plugin on GitHub (pfnet-research)
- Selected for Japan Digital Agency's government AI environment ("Gennai")
ASEAN Perspective
PLaMo (Preferred Networks) in Southeast Asia
ASEAN-region availability and pricing notes coming soon. Drop the editorial team a note via /contact/ if you can supply local context (Singapore/Malaysia/Indonesia/Thailand/Vietnam).
PLaMo is one of the few Japanese LLM efforts built genuinely from scratch rather than fine-tuned on a Western base, and it's now backed by government validation: the Digital Agency rolled PLaMo Translate into its internal "Gennai" AI environment in December 2025, and by March 2026 PLaMo 2.0 Prime was selected as a trial model for the government's broader AI initiative. PLaMo 2.0 Prime also won the 2025 Nikkei Excellent Products and Services Award grand prize — an unusually strong signal for a domestic model competing against OpenAI and Anthropic's Japan-market presence.
PLaMo 3.0 Prime shipped officially on June 22, 2026, alongside PLaMo Lite, PLaMo Translate and a finance-tuned variant — a coherent lineup rather than a single model. API pricing (300 yen per million input tokens, 1,000 yen per million output tokens) undercuts most comparable-quality options for Japanese-language work. The catch: documentation, sign-up and support are Japanese-only, and open-weight versions carry a Community or Non-Commercial license with a 1-billion-yen revenue cap for free commercial use — this is built for Japan, not a global or ASEAN audience.
What people say
PLaMo's biggest 2026 story is government adoption. Japan's Digital Agency began internal use of PLaMo Translate inside its generative-AI environment, codenamed "Gennai," in December 2025, with rollout to other ministries and agencies planned through 2026. Then in March 2026, PFN announced PLaMo 2.0 Prime was selected as a trial model for the Digital Agency's Government AI project — a rare case of a Japanese-built foundation model getting direct public-sector validation rather than just enterprise pilots. Separately, PLaMo 2.0 Prime won the grand prize at the 2025 Nikkei Excellent Products and Services Awards.
The product lineup expanded through the year: PLaMo 3.0 Prime launched officially on June 22, 2026 as PFN's flagship domestically-built foundation model, positioned alongside PLaMo Lite (edge/lightweight deployment), PLaMo Translate (Japanese-specialized translation) and a finance-tuned variant with additional domain training. PLaMo API is OpenAI-API-compatible, so existing code using openai-python or LangChain integrates with minimal changes — a practical detail for teams evaluating a swap-in domestic alternative. Launch pricing for PLaMo Prime API was set at 300 yen per million input tokens and 1,000 yen per million output tokens, with a 1,000-yen credit for new signups during the launch campaign — competitive against comparable Japanese-capable models.
Open-weight releases sit on Hugging Face under the pfnet organization, spanning small edge models (1B, 2B) up to PLaMo-100B, plus newer PLaMo 3 NICT variants (2B/8B/31B) built in partnership with Japan's National Institute of Information and Communications Technology. Licensing varies by version: PLaMo-100B uses a Non-Commercial license, while PLaMo 3 NICT models fall under the PLaMo Community License, which permits commercial use only after registering with PFN and only if the user's (or affiliated company's) annual revenue stays under 1 billion yen — a meaningful gate for larger commercial deployments. GitHub hosts implementation tooling (a vLLM inference plugin) rather than the model weights themselves, under the pfnet-research organization, which lists over 100 repositories overall.
Summary of public user & expert reviews, compiled by RECATOOLS.
About this listing
This entry was compiled from publicly available data including PLaMo (Preferred Networks)'s official website, press releases, documentation, and reputable third-party publications. RECATOOLS is not affiliated with PLaMo (Preferred Networks) unless explicitly stated.
Third-party AI tools update their pricing, features, availability, and policies frequently. Information here may be outdated by the time you read this — we make reasonable efforts to keep listings current, but cannot guarantee absolute accuracy.
For the latest details, please refer to PLaMo (Preferred Networks) directly →
Spotted something out of date? Suggest an update →
Alternatives to PLaMo (Preferred Networks)
More in LLMs & Chat
