February 11, IBM has announced that DeepSeek-R1 distillation versions of Llama 3.1 8B and Llama 3.3 70B are now available on IBM's enterprise AI development platform, watsonx.ai. According to the official introduction, DeepSeek has also leveraged knowledge distillation to optimize multiple Llama and Qwen variants with data generated from the R1 model. Qwen variants using data generated from the R1 model. Users can use DeepSeek distillation models on watsonx.ai by...

IIBM has announced new updates to its generative AI platform watsonx, including new assistant tools, third-party models and a broader commitment to open source AI. At the company's annual Think conference, IBM announced a shift in focus to support open source. The company has open-sourced a series of its Granite large-scale language models, meaning that organizations can use them to support their business applications. These new open source models include powerful coding models that developers can use to troubleshoot, generate code, and maintain code bases across 116 programming languages.Gr...