NVDA 216.47 ▼0.29%GOOGL 340.28 ▼1.13%MSFT 531.87 ▼0.31%AMD 534.74 ▼0.62%INTC 91.69 ▼0.37%TSMC 425.08 ▲1.16%AMZN 262.63 ▲0.07%META 694.25 ▼1.59%AAPL 339.50 ▼1.00%PLTR 192.60 ▼0.80%
Markets at last close

Models

GoTo opens language model trained for Indonesia

·1 min read

GoTo released a 27 billion parameter open-source language model at Bali Tech Week on 19 October 2026. Trained for Bahasa Indonesia, Javanese, Sundanese, Balinese and Bataknese, it is freely available for startups, businesses, universities, researchers and public institutions to use, study and adapt. The release builds on open-model work the company began in 2024.

Training material included Indonesian-language text from media partners and public-sector websites, alongside regional-language data. GoTo collaborated with Universitas Sumatera Utara to collect Bataknese data reflecting contemporary usage. The company describes the model as capable of demanding language tasks while remaining small enough for universities, startups and public institutions to run on their own servers rather than in a large data center.

GoTo CEO Hans Patuwo said locally developed AI can address Indonesian data and regional-language needs that global tools may not fully serve. The company is also developing local technical skills, with interns from seven universities working on data preparation, model training and testing. Interns contributed voice recordings for Indonesian speech technology and helped collect Batak-language data. The model is available through Hugging Face.

Originally reported by gotocompany.comRead the source →
Related coverage