The translation race is no longer silicon valley's to lose. japan's preferred networks released…
the translation race is no longer silicon valley's to lose. japan's preferred networks released plamo 3 translate, a 31b-parameter model, on oct 8, aimed at professional-grade japanese translation that competes with the frontier labs.
the language layer is going regional.
Context
Preferred Networks announced plamo-3-translate on October 7, a translation-specialised model that succeeds plamo-2-translate and now powers its PLaMo Translate service. PFN's own post says it spent hundreds of times more training tokens and data-synthesis compute than the previous generation.
PFN positions it as matching overseas frontier models on Japanese translation at much lower cost. Those are the vendor's own claims.
The 31b figure: PFN's public 31B model is the PLaMo 3 NICT base on Hugging Face. The translate model's parameter count is not stated in the sources read, so it is unsupported here, not refuted.
Date: PFN's post is dated October 7 and Japanese press covered it October 8. The note's Oct 8 is the press day.
Frontier-level quality is PFN's claim. No independent benchmark was found.
Related work
- PFN: plamo-3-translate announcement ↗The vendor post with the training-compute claim.
- PLaMo Translate service page ↗Where the model is served.
- PLaMo 3 NICT 31B base on Hugging Face ↗The 31B base model, built with NICT.
Watch next
- Independent Japanese-English translation benchmarks for plamo-3-translate once published.
Sources
- PFN: plamo-3-translate (Japanese)preferred.jp
- Impress Watch coverage, Oct 8, 2026forest.watch.impress.co.jp
- PLaMo Translatetranslate.preferredai.jp
- Hugging Face: plamo-3-nict-31b-basehuggingface.co
Provenance
The note above is reproduced unedited from the original post, first published on Threads on 9 October 2026 at 17:03 IST. Sources are the papers and datasets the note draws on.
View the original post ↗Embed this note
More notes
The air is now being asked to keep its own ledger
the air is now being asked to keep its own ledger: ecmwf’s aifs compo becomes the first ai model to forecast atmospheric composition globally every three hours, cleanair simulates 365 days of pm2.5 over china in ten seconds, and a unified framework maps six pollutants at one kilometer across the whole country. the air now files its own composition report.
read the note →The current is now being asked to draw its own map
the current is now being asked to draw its own map: china’s langya 2.0 predicts six ocean phenomena including internal waves and mesoscale eddies, a deep net called wenhai resolves eddies globally with air sea flux formulas built in, and scripps infers surface currents from the way temperature patterns deform in satellite images. the ocean now files its own circulation report.
read the note →The soil is now being asked to report its own carbon
the soil is now being asked to report its own carbon: a nix color sensor paired with generative data augmentation predicts soil organic carbon without a lab, random forest drives 74 percent of soil health mapping studies, and sentinel 2 tracks five year carbon change across france and italy from 922 samples. the dirt now files its own carbon account.
read the note →