A new trillion-parameter model, providing super strong question-answering and text generation capabilities.

Based on the yi-large model, supports and enhances tool invocation capabilities, suitable for various business scenarios requiring agent or workflow construction.

High-level service based on the yi-large super strong model, combining retrieval and generation techniques to provide precise answers and real-time information retrieval services.

Exceptional performance at a high cost-performance ratio. Conducts high-precision tuning based on performance, inference speed, and cost.

Medium-sized model upgraded and fine-tuned, balanced capabilities, and high cost-performance ratio. Deeply optimized instruction-following capabilities.

200K ultra-long context window, providing deep understanding and generation capabilities for long texts.

Small yet powerful, lightweight and fast model. Provides enhanced mathematical computation and coding capabilities.

Model for complex visual tasks, providing high-performance image understanding and analysis capabilities.

Initial version, recommended to use yi-large (new version).