Skip to main content
Back to News
Preferred Networks (PFN) has launched its flagship domestic large language model, PLaMo 3.0 Prime, o...
Product
2 min read
JP

Preferred Networks (PFN) has launched its flagship domestic large language model, PLaMo 3.0 Prime, o...

The AMW Read

Incremental distribution update for a known player (PFN) extending its sovereign AI footprint in Japan, with segment-level implications for enterprise adoption.
NoveltySignificance
Foundation Models · Player Map

Preferred Networks (PFN) has launched its flagship domestic large language model, PLaMo 3.0 Prime, on Sakura Internet's AI inference API platform, Sakura's AI Engine, starting August 4. The deployment marks the first availability of PFN's full-scratch reasoning model through a Japanese cloud provider's API, with access granted on an application basis (free tier excluded). The model, based on pre-training research with NICT, extends context length from 64k to 256k tokens, enhances Japanese language performance, and claims competitive results in instruction following, coding, and tool use against open models like Qwen3.6-27B and closed models like GPT-5.4 mini and Claude Haiku 4.5. Pricing is disclosed only to approved users.

This partnership underscores a strategic move by PFN to distribute its models through domestic cloud infrastructure, capitalizing on Japan's push for sovereign AI capabilities. By leveraging Sakura's high-performance data centers, PFN gains a distribution channel that appeals to enterprises prioritizing data residency and regulatory compliance. For the broader market, PLaMo 3.0 Prime's availability on an API platform lowers the barrier for Japanese enterprises to adopt a locally developed, reasoning-capable LLM, potentially accelerating AI integration in sectors like finance and government where data sovereignty is critical. It also intensifies competition among Japanese cloud providers to host preferred domestic models, aligning with national interests in technological self-reliance.

For builders, this means access to a frontier-scale reasoning model with strong Japanese-language support and extended context, ideal for agentic workflows and complex document processing. However, the application-only access and non-public pricing could create friction for rapid prototyping. Investors should view this as a signal of PFN's go-to-market maturation, moving from research to productized distribution via strategic partnerships—a pattern that may increase PFN's enterprise traction and valuation potential. The success of this deployment could set a precedent for other domestic AI labs to partner with local cloud providers, reshaping the Japanese AI infrastructure landscape.

#PreferredNetworks #PLaMo #SakuraInternet #JapaneseAI #SovereignAI #LLM

#Preferred Networks#PLaMo 3.0 Prime#Sakura Internet#AI inference#sovereign AI#related:Sakura Internet

How This Connects

Based on Foundation Models · Player Map

  1. 20h agoDeepSeek, the Chinese AI lab known for its low-cost models, has announced significant API price incr...DeepSeek
  2. 1d agoApple reportedly trains China-specific LLM with Alibaba, pursuing dual AI strategyApple
  3. 1d agoAlibaba's Qwen team has open-sourced Qwen3.8-27B, a 27-billion-parameter multimodal model designed f...Qwen
  4. 1d agoAlibaba has released Qwen 3.8 27B, an Apache 2.0-licensed open-weight dense model with 27 billion pa...Alibaba Qwen 3.8 27B launch
  5. 2d agoZhipu AI (智谱) released GLM-5.3, a new open-weight foundation model with advanced cybersecurity capab...Z.ai
  6. 1w agoPreferred Networks (PFN) has launched its flagship domestic large language model, PLaMo 3.0 Prime, o... · THIS ARTICLE

Related News

More news from Preferred Networks

Stay updated with the latest news and announcements from Preferred Networks.

View all Preferred Networks news

Discover AI Startups

Explore 5,000+ AI companies with VC-grade analysis, funding data, and investment insights.

Explore Dashboard