The AllSpark team has introduced two new open-source search agents, Iris-mini and Iris-pro, both built on the Qwen model architecture. These agents have set new benchmarks among open-weight models within their respective size categories.
According to The Decoder, Iris-mini and Iris-pro outperform other open-weight search agents in their class. The training data and model design have enhanced their capabilities, enabling improved performance on tasks beyond their original training scope, including general tool use and office-related work.
For Japanese markets, where automation and AI integration are accelerating, these developments in open-source AI tools could provide valuable resources for enhancing productivity in finance, crypto, and equities sectors.