PFN Unveils PLaMo 3 Translate 31B, Matching Top Giants in Translation
Preferred Networks (PFN), a leading Japanese enterprise, officially unveiled its latest translation model, PLaMo3Translate31B, yesterday. This release highlights exceptional proficiency in Japanese translation while addressing multilingual requirements, marking it as a fully indigenous large-scale model developed in Japan. The training dataset comprises approximately 30% Japanese content.

Technically, PLaMo3Translate31B is built on the PLaMo3 foundation. Expanding beyond the previous generation’s support for just Japanese and English, this model now supports 53 languages and introduces an innovative online meeting translation feature for 15 of them.
Regarding performance and cost efficiency, PFN reports that PLaMo3Translate31B surpasses OpenAI’s GPT-6.1Sol and GPT-6Astra in average scores across four authoritative translation benchmarks. Its optimized processing for Japanese results in highly competitive operational costs, priced at just 14 yen (approximately 0.59 Chinese yuan) per 100,000 characters, significantly undercutting comparable market offerings.
Related article
Tesla Hypes In-Car AI Voice With Dual Domestic Large Models
Tesla’s in-car voice interaction in China is poised for a major overhaul. The official Tesla China account recently unveiled a promotional video highlighting "voice assistant control," demonstrating the updated system’s capabilities and new features.
New EV Startup Claims AI Infotainment Renders CarPlay Screen Mirroring Obsolete
The rise of intelligent vehicle interfaces is rendering traditional smartphone projection systems increasingly irrelevant. Waseem Bensaid, Rivian’s software chief, recently stated that with rapid AI advancements, integrating Apple CarPlay has become
Why OpenAI’s Head of Ethics Chloé Bakalar Departed
Chloe Bakalar, former Head of Ethics at OpenAI. Credit: CITP Princeton/YouTubeOpenAI’s Ethics Head Chloé Bakalar has left the leading AI amid a tumultuous period for AI safety – following the OpenAI hacking incidentOpenAI’s Head of EthicsChloé Bakala
Related Special Topic Recommendations
Comments (0)
0/500
Preferred Networks (PFN), a leading Japanese enterprise, officially unveiled its latest translation model, PLaMo3Translate31B, yesterday. This release highlights exceptional proficiency in Japanese translation while addressing multilingual requirements, marking it as a fully indigenous large-scale model developed in Japan. The training dataset comprises approximately 30% Japanese content.

Technically, PLaMo3Translate31B is built on the PLaMo3 foundation. Expanding beyond the previous generation’s support for just Japanese and English, this model now supports 53 languages and introduces an innovative online meeting translation feature for 15 of them.
Regarding performance and cost efficiency, PFN reports that PLaMo3Translate31B surpasses OpenAI’s GPT-6.1Sol and GPT-6Astra in average scores across four authoritative translation benchmarks. Its optimized processing for Japanese results in highly competitive operational costs, priced at just 14 yen (approximately 0.59 Chinese yuan) per 100,000 characters, significantly undercutting comparable market offerings.
Tesla Hypes In-Car AI Voice With Dual Domestic Large Models
Tesla’s in-car voice interaction in China is poised for a major overhaul. The official Tesla China account recently unveiled a promotional video highlighting "voice assistant control," demonstrating the updated system’s capabilities and new features.
New EV Startup Claims AI Infotainment Renders CarPlay Screen Mirroring Obsolete
The rise of intelligent vehicle interfaces is rendering traditional smartphone projection systems increasingly irrelevant. Waseem Bensaid, Rivian’s software chief, recently stated that with rapid AI advancements, integrating Apple CarPlay has become
Why OpenAI’s Head of Ethics Chloé Bakalar Departed
Chloe Bakalar, former Head of Ethics at OpenAI. Credit: CITP Princeton/YouTubeOpenAI’s Ethics Head Chloé Bakalar has left the leading AI amid a tumultuous period for AI safety – following the OpenAI hacking incidentOpenAI’s Head of EthicsChloé Bakala





Home






