IT Home Releases Ling-3.1-flash Model with 560 Billion Parameters
IT Home has launched the Ling-3.1-flash AI model, featuring approximately 560 billion total parameters and 25 billion active parameters per token. The model boasts a 1 million token context window.
Chinese technology firm IT Home has unveiled its latest AI model, Ling-3.1-flash. The model reportedly contains around 560 billion total parameters, with approximately 25 billion parameters activated per token. A key feature is its extensive 1 million token context window, designed to handle longer documents, code, and task histories.
The development focused on optimizing the model for tasks such as general intelligent agents, search functionalities, daily office work, and software development. IT Home also stated efforts to enhance its capabilities in fields like healthcare, finance, and materials science research.
Ling-3.1-flash will be available for a two-week free trial period. During this trial, the service length will be limited to 256,000 tokens. Following the trial, when the service transitions to a paid model, the full 1 million token context will be accessible.
IT Home also plans to release the model as open source after the trial period. The company intends to continue updating and improving the model's performance and expand its applications.