GoTo Unveils Open-Source AI Model Built for Indonesia

A textured flat illustration in a warm, cozy editorial style. A small sapling growing from a patch of ground patterned like woven batik cloth, its leaves rendered as simple circuit-line shapes, symbolising a locally grown AI model rooted in Indonesian culture and language. Style: Simplified, stylized forms with minimal detail. Soft organic grain texture. Muted autumnal color palette (warm yellows, oranges, browns, sage greens, cream backgrounds). Single soft light source creating subtle shadows. Flat shapes with no 3D rendering. Thin consistent outlines. Mid-century modern illustration influence. Composition: Clear focal point, balanced layout, environmental details rendered simply. Generous negative space. 16:9 landscape. Mood: Calm, warm, relatable, editorial, lifestyle. Technical: High resolution, matte finish, no harsh gradients. No text, no logos, no recognisable faces.

Indonesian super-app group GoTo has released a 27-billion-parameter open-source large language model trained specifically for Bahasa Indonesia and four regional languages, in a push to give the country’s AI developers a locally grounded starting point.

The model, unveiled at Bali Tech Week in Sanur on 19 October, was trained on Bahasa Indonesia alongside Javanese, Sundanese, Balinese and Batak. It builds on open-model work GoTo began in 2024 and is now freely available for startups, businesses, universities, researchers and public institutions to use, study and adapt.

Built on local data

The model was trained on Indonesian-language text drawn from media partners and public-sector websites, alongside additional regional-language datasets. For Batak, a language with far less digital text available than Bahasa Indonesia, GoTo worked with Universitas Sumatera Utara to collect data reflecting how the language is actually used today.

At 27 billion parameters, the model sits in the mid-size range — large enough for demanding language tasks such as answering questions, summarising documents and generating translations, but small enough for a university, startup or public institution to run on its own servers rather than relying on a large data centre.

“Indonesia makes great use of global AI, but we must build our own too, to solve the problems that are unique to us. AI created for global use may not be fully optimised for our localised data or the regional languages so many Indonesians grow up using at home,” said Hans Patuwo, Group CEO of GoTo.

“At GoTo, we are overcoming that challenge by building locally relevant AI for use with our own customers. Opening this technology to everyone is our way of giving back to the country, providing anyone building AI in Indonesia with a starting point rooted in the local context,” he added.

Part of a national AI push

The release supports the Ministry of Communication and Digital Affairs’ Indonesia Digital Vision 2045, which aims to build an inclusive, innovation-driven digital ecosystem and move the country from being a consumer of AI technology to a creator of it.

GoTo is also putting resources into training the people who will build Indonesia’s next generation of AI systems. Interns from seven universities — stretching from Sumatra to Sulawesi, including Udayana University in Bali — worked alongside GoTo’s AI teams on data preparation, model training and testing. Some also lent their voices to train the company’s Indonesian speech technology and helped gather Batak-language data.

The model is available for download on Hugging Face.

Author


Discover more from techcoffeehouse.com

Subscribe to get the latest posts sent to your email.

Use promo code “TCH15” to get 15% off on checkout.

Share your thoughts

This site uses Akismet to reduce spam. Learn how your comment data is processed.

Discover more from techcoffeehouse.com

Subscribe now to keep reading and get access to the full archive.

Continue reading

Discover more from techcoffeehouse.com

Subscribe now to keep reading and get access to the full archive.

Continue reading