SuperX, formerly known as Junee Limited, has officially rebranded and shifted its focus to AI infrastructure, adopting the NASDAQ ticker symbol 'SUPX'.

The company has launched its All-in-One Multi-Model Servers (MMS), designed to facilitate the dynamic collaboration of multiple AI models, providing enterprises with secure and efficient AI solutions.

These MMS servers come pre-configured with OpenAI's high-performance large language models, GPT-OSS-120B and GPT-OSS-20B, which have been shown to outperform several leading closed-source models in key benchmarks.

Benchmark tests indicate that the GPT-OSS-120B model excels in areas such as language understanding and mathematics exams, confirming its competitive edge over closed-source alternatives.

Key features of the MMS include advanced data security through NVIDIA's Confidential Computing technology, rapid deployment capabilities, and support for over 60 pre-configured scenario-based agents.

Kenny Sng has been appointed as the new Chief Technology Officer of SuperX, emphasizing the importance of multi-model collaboration in advancing AI technology toward Artificial General Intelligence (AGI).

With pre-configured templates and low-code interfaces, business users can quickly build intelligent agents for various enterprise applications, enhancing workflow efficiency.

The MMS also features a built-in knowledge base, cloud-coordinated model caching for instant access to the latest models, and comprehensive integration of hardware and software.

The MMS series is available in different editions, catering to various enterprise sizes, with pricing starting from $50,000 for individual workstations to $4 million for cluster editions.

These servers ensure data security and can be deployed without extensive IT resources, making them accessible for enterprises of all sizes.

SuperX's integration of high-performance open-source models provides a competitive edge, matching or exceeding some leading closed-source models in performance benchmarks.

The incorporation of NVIDIA Confidential Computing technology not only ensures data security but also protects AI intellectual property during training and inference.