Project AURIONAI Use-Cases Research & Innovation
LLM

Llama-Sahabat-AI-v2-70B-IT

Indonesian 70B instruct model by GoTo and Indosat, developed with AI Singapore; evaluated on IndoMMLU, SEA-HELM and SEA-IFEval (scores not listed on card).

GoTo Gojek TokopediaGoTo Gojek TokopediaIndonesiaIndosat Ooredoo HutchisonIndosat Ooredoo HutchisonIndonesiaAI SingaporeAI SingaporeSingaporeSahabat-AIFamily
GoTo Gojek Tokopedia
Levelmodel
ClassLLM
Parameters71B
Contextnot in the notes
Released2 Jun 2025
Opennessopen weights restricted
Runs onmulti-GPU node
Statusactive

Lineage

continued pretrainfinetune
Llama 3.1 70B Instruct to Llama-Sahabat-AI-v2-70B-IT: continued pretrain, medium confidenceLlama 3.1 70B Instruct to Llama-3.1-Nemotron-70B-Instruct: finetune, high confidenceLlama 3.1 70B to Llama-Sahabat-AI-v2-70B-IT: finetune, high confidenceLlama 3.1 70B InstructLlama 3.1 70B InstructMetaLlama 3.1 70BLlama 3.1 70BMetaLlama-3.1-Nemotron-70B-InstructLlama-3.1-Nemotron-70B-Ins…NVIDIA, Oct 2024Llama-Sahabat-AI-v2-70B-ITLlama-Sahabat-AI-v2-70B-ITGoTo Gojek Tokopedia, Jun 2025
MetaLlama 3.1 70B Instruct is the parent (continued pretrain)medium confidencenotes
MetaLlama 3.1 70B is the parent (finetune)high confidenceHugging Face base_model
Where this comes from
Parent and child relations from model cards, Hugging Face base_model fields and the genealogy tables of the research notes. The tree shows two generations each way, 4 models. Confidence and evidence are the notes' own.

Industry Use Cases

Finance
Banking customer service assistants
Public Sector
Indonesian and regional-language LLMs

R&D Topics

Regional or national foundation model training (continued pretraining at 7B to 70B)exemplarLow-resource and regional language resourcesproduced by

Identity

Developer
GoTo Gojek Tokopedia, Indosat Ooredoo Hutchison, AI Singapore
Family
Sahabat-AI
Released
2 Jun 2025 sourcesregistry date
Weights published
–
Record checked
10 Oct 2026
Where this comes from
Developer and dates from the research notes. Release date is the vendor's public announcement; weights date is when open weights were published.

Architecture and Size

Architecture
not in the notes
Total parameters
71B sources
Active parameters
–
Checkpoint parameters
–
Note
70B (per name); no base_model declared (relation finetune)
Context length
not in the notes
Input
text
Output
text
Where this comes from
Parameter counts as the developer states them. Where the checkpoint holds a different total, both are shown (register A16).

Openness and Access

Openness
open weights restricted
Licence
Llama 3.1 Community License sources
Access routes
open weights download
Where this comes from
The licence file governs over registry tags and aggregator labels.

Deployment

Runs on
multi-GPU node
Footprint
About 140 GB in BF16
Runtimes
not in the notes
Where this comes from
Deployment tiers and runtimes as named in the notes. Footprint is the notes' figure, not a measurement by AURION.

Indonesia

Indonesian support
listed by developer
  • IndoMMLU and SEA-HELM (model card says both were run), run by GoTo / Sahabat-AI team (developer-reported) huggingface.co
Regional languages
Javanese, Sundanese, Batak Toba, Balinese
Jakarta hosting
unknown
Languages
English, Indonesian, Javanese, Sundanese, Batak Toba, Balinese
Where this comes from
Indonesia fields are carried on every model; unknown is an allowed value.

Benchmark Results

The notes record no benchmark result for this model.

Sources

  1. GoToCompany (PT GoTo Gojek Tokopedia Tbk) huggingface.co https://huggingface.co/GoToCompany
  2. GoToCompany/Llama-Sahabat-AI-v2-70B-IT , Hugging Face huggingface.co https://huggingface.co/GoToCompany/Llama-Sahabat-AI-v2-70B-IT
  3. huggingface.co/api/models/GoToCompany/Llama-Sahabat-AI-v2-70... https://huggingface.co/api/models/GoToCompany/Llama-Sahabat-AI-v2-70B-IT
  4. Massive enhancement announced for Indonesian LLM - Developing Telecoms developingtelecoms.com https://developingtelecoms.com/telecom-technology/enterprise-ecosystems/18569-massive-enhancement-announced-for-indonesian-llm.html
  5. Sahabat-AI Gets Smarter: Indosat and GoTo Launch New 70 Billion Parameter Model with Multilingual Chat Service gotocompany.com https://www.gotocompany.com/en/news/press/sahabat-ai-gets-smarter-indosat-and-goto-launch-new-70-billion-parameter-model-with-multilingual-chat-service

To cite this page: Project AURION, AI Use-Cases Research & Innovation, "Llama-Sahabat-AI-v2-70B-IT", https://aurion.semai.id/models/goto-llama-sahabat-ai-v2-70b-it/, records of October 2026.