Despite growing international scrutiny of the Hangzhou-based company, which claims it can develop industry-leading models at a fraction of the usual cost, US technology firms have been rushing to adopt the latest artificial intelligence ( AI ) reasoning model from Chinese start-up DeepSeek.
Advertisement
Major chip designer Nvidia has made DeepSeek’s R1 unit available to users of its NIM middleware since Thursday, saying the design provides” state-of-the-art reasoning features”, “high inference efficiency”, as well as “leading accuracy” for jobs requiring logical inference, reasoning, mathematics, coding and vocabulary understanding.
DeepSeek’s open-source logic design, R1, released on January 20, has shown features similar to OpenAI’s closed-source GPT models in specific areas, but at considerably lower education costs.
Advertisement