Tower+: Bridging Generality and Translation Specialization in Multilingual LLMs
Rei, Ricardo, Guerreiro, Nuno M., Pombal, José, Alves, João, Teixeirinha, Pedro, Farajian, Amin, Martins, André F. T.
–arXiv.org Artificial Intelligence
Fine-tuning pretrained LLMs has been shown to be an effective strategy for reaching state-of-the-art performance on specific tasks like machine translation. However, this process of adaptation often implies sacrificing general-purpose capabilities, such as conversational reasoning and instruction-following, hampering the utility of the system in real-world applications that require a mixture of skills. In this paper, we introduce Tower+, a suite of models designed to deliver strong performance across both translation and multilingual general-purpose text capabilities. We achieve a Pareto frontier between translation specialization and multilingual general-purpose capabilities by introducing a novel training recipe that builds on Tower (Alves et al., 2024), comprising continued pretraining, supervised fine-tuning, preference optimization, and reinforcement learning with verifiable rewards. At each stage of training, we carefully generate and curate data to strengthen performance on translation as well as general-purpose tasks involving code generation, mathematics problem solving, and general instruction-following. We develop models at multiple scales: 2B, 9B, and 72B. Our smaller models often outperform larger general-purpose open-weight and proprietary LLMs (e.g., Llama 3.3 70B, GPT-4o). Our largest model delivers best-in-class translation performance for high-resource languages and top results in multilingual Arena Hard evaluations and in IF-MT, a benchmark we introduce for evaluating both translation and instruction-following. Our findings highlight that it is possible to rival frontier models in general capabilities, while optimizing for specific business domains, such as translation and localization.
arXiv.org Artificial Intelligence
Jun-23-2025
- Country:
- South America (0.04)
- North America
- Central America (0.04)
- United States > Florida
- Miami-Dade County > Miami (0.04)
- Europe
- Sweden > Stockholm
- Stockholm (0.04)
- Portugal > Lisbon
- Lisbon (0.14)
- Middle East > Republic of Türkiye
- Istanbul Province > Istanbul (0.04)
- Sweden > Stockholm
- Asia
- Singapore (0.04)
- Thailand > Bangkok
- Bangkok (0.04)
- Middle East
- UAE > Abu Dhabi Emirate
- Abu Dhabi (0.04)
- Republic of Türkiye > Istanbul Province
- Istanbul (0.04)
- UAE > Abu Dhabi Emirate
- Genre:
- Research Report > New Finding (0.34)
- Industry:
- Law (0.46)
- Technology: