SEA-LION v4 (Latest)

SEA-LION version 4, released on Aug 25, is our first multimodal models trained on Southeast Asian text, each with their unique strengths:

• Gemma-SEA-LION-v4-27B is suited for translation, abstractive summarisation, natural language inference and causal reasoning, metaphor understanding, question answering, paraphrase and sentiment analysis applications where regional language support and advanced reasoning is critical.

• Gemma-SEA-LION-v4-27B-IT is suited for knowledge-intensive tasks, and for high-demand contexts where comprehensive language comprehension are essential.

• Gemma-SEA-LION-v4-27B-IT-GGUF, NVFP4, FP8_Dynamic supports inference on a range of consumer-grade GPUs and are compatible with various inference engines.

• Gemma-SEA-LION-v4-27B-VL is our powerful vision-language model, expertly trained to interpret both images and text with a deep, nuanced understanding of Southeast Asian cultural contexts.

• Qwen-SEA-LION-v4-32B-IT is our flagship instruction-tuned model, designed for maximum performance in general SEA context tasks.

• Qwen-SEA-LION-v4-32B-IT-4BIT, 8BIT offers a near-perfect balance of performance and efficiency, making the power of SEA-LION accessible on resource-constrained hardware like consumer-grade GPUs.

SEA-LION v4 continues our mission to create language models that understand and respond with greater cultural awareness and depth across Southeast Asia. In addition, SEA-LION v4 has the ability to handle both image and text input as a multimodal model.

For detailed information of each of the SEA-LION v4 models, please refer to their individual documentation pages via the links above.