Colombo, Sri Lanka
Sri Lanka’s growing artificial intelligence ecosystem achieved a significant milestone this week with the official public release of Chat2Find-Instruct-v1, a powerful trilingual large language model (LLM) developed to support Sinhala, Tamil, and English. The model is now available globally through the widely used AI platform Hugging Face, while local access and distribution are being facilitated through Lanka Data, strengthening Sri Lanka’s position in the rapidly evolving AI landscape.
Built upon the advanced Qwen architecture and fine-tuned using high-quality multilingual datasets, Chat2Find-Instruct-v1 has been specifically engineered for reasoning, instruction following, and agentic AI workflows. The model is designed to process complex logical tasks, execute tool-calling functions, and provide high-quality responses across Sri Lanka’s three primary languages.
The release represents one of the most significant advancements in Sri Lankan language AI to date. While many global AI models offer limited support for Sinhala and Tamil, Chat2Find-Instruct-v1 was trained with a strong emphasis on native language understanding, ensuring more accurate and culturally relevant responses for local users.
According to technical documentation released alongside the model, the training dataset consisted of more than 279,000 curated records, with Sinhala and Tamil content accounting for over 80 percent of the training material. This focus enables the model to better understand local contexts, language nuances, and multilingual interactions that are often overlooked by international AI systems.
A standout feature of Chat2Find-Instruct-v1 is its advanced chain-of-thought reasoning capability, allowing the model to process complex problems step-by-step before generating responses. The system also supports agentic tool calling, enabling integration with external APIs, databases, and enterprise workflows.
Industry observers note that the public release of a locally developed trilingual reasoning model could have significant implications for education, legal technology, customer service, government digitisation, and enterprise AI adoption throughout Sri Lanka.
The model’s availability through Hugging Face provides immediate access to researchers, developers, and organisations worldwide, while Lanka Data’s involvement is expected to accelerate local adoption and create opportunities for Sri Lankan businesses seeking AI solutions tailored to regional languages.
AI experts view the launch as a landmark achievement for Sri Lanka’s technology sector, demonstrating that locally trained and culturally aligned language models can compete in an increasingly global AI market.
As governments and enterprises across South Asia seek AI systems capable of understanding regional languages and contexts, Chat2Find-Instruct-v1 may serve as an important example of how emerging economies can build their own sovereign AI capabilities rather than relying solely on foreign-developed models.
The release also highlights the growing maturity of Sri Lanka’s AI research community and signals the country’s ambitions to become a regional hub for multilingual artificial intelligence innovation.
The model is now publicly accessible through Hugging Face and is expected to undergo further enhancements as developers and researchers contribute feedback and new use cases. With its focus on reasoning, multilingual understanding, and practical AI deployment, Chat2Find-Instruct-v1 marks a significant step forward in Sri Lanka’s journey toward AI independence and technological innovation.
