# Sabi'a-3 Technical Report

arXiv: https://arxiv.org/abs/2410.12049 · HTML: https://mlsanity.com/paper/2410.12049/

## Key facts

- Cited by 7 papers in the map (data through 14 Oct 2026).
- Builds on 7 of its 13 references that are in the map.
- First posted on arXiv 15 Oct 2024.

## Builds on (most cited first, 7 of 7)

- [Judging LLM-as-a-Judge with MT-Bench and Chatbot Arena](https://mlsanity.com/paper/2306.05685/) (2023)
- [Phi-3 Technical Report: A Highly Capable Language Model Locally on Your Phone](https://mlsanity.com/paper/2404.14219/) (2024)
- [Instruction-Following Evaluation for Large Language Models](https://mlsanity.com/paper/2311.07911/) (2023)
- [Yi: Open Foundation Models by 01.AI](https://mlsanity.com/paper/2403.04652/) (2024)
- [CodeAgent: Enhancing Code Generation with Tool-Integrated Agent Systems for Real-World Repo-level Coding Challenges](https://mlsanity.com/paper/2401.07339/) (2024)
- [QuALITY: Question Answering with Long Input Texts, Yes!](https://mlsanity.com/paper/2112.08608/) (2021)
- [Taken out of context: On measuring situational awareness in LLMs](https://mlsanity.com/paper/2309.00667/) (2023)

## Built on by (most cited first, 7 of 7)

- [TiEBe: Tracking Language Model Recall of Notable Worldwide Events Through Time](https://mlsanity.com/paper/2501.07482/) (2025)
- [Ticket-Bench: A Kickoff for Multilingual and Regionalized Agent Evaluation](https://mlsanity.com/paper/2509.14477/) (2025)
- [PoETa v2: Toward More Robust Evaluation of Large Language Models in Portuguese](https://mlsanity.com/paper/2511.17808/) (2025)
- [BRoverbs -- Measuring how much LLMs understand Portuguese proverbs](https://mlsanity.com/paper/2509.08960/) (2025)
- [JurisTCU: A Brazilian Portuguese Information Retrieval Dataset with Query Relevance Judgments](https://mlsanity.com/paper/2503.08379/) (2025)
- [Aplicacc~ao de Large Language Models na An'alise e S'intese de Documentos Jur'idicos: Uma Revis~ao de Literatura](https://mlsanity.com/paper/2504.00725/) (2025)
- [CAPITU: A Benchmark for Evaluating Instruction-Following in Brazilian Portuguese with Literary Context](https://mlsanity.com/paper/2603.22576/) (2026)

## Abstract (arXiv)

This report presents Sabi'a-3, our new flagship language model, and Sabiazinho-3, a more cost-effective sibling. The models were trained on a large brazilian-centric corpus. Evaluations across diverse professional and academic benchmarks show a strong performance on Portuguese and Brazil-related tasks. Sabi'a-3 shows large improvements in comparison to our previous best of model, Sabia-2 Medium, especially in reasoning-intensive tasks. Notably, Sabi'a-3's average performance matches frontier LLMs, while it is offered at a three to four times lower cost per token, reinforcing the benefits of domain specialization.