Multiverse Computing Unveils Free Compressed AI Model, Pioneering Cost-Effective Solutions
In the rapidly evolving landscape of artificial intelligence, the size and computational demands of large language models (LLMs) have posed significant challenges for widespread adoption. Addressing this issue head-on, Spanish startup Multiverse Computing has introduced a groundbreaking solution: compressed AI models that maintain high performance while significantly reducing resource requirements.
At the heart of this innovation is CompactifAI, a compression technology inspired by quantum computing principles. This proprietary technology enables the downsizing of extensive AI models without compromising their efficacy. By applying CompactifAI to models developed by OpenAI, Multiverse Computing has successfully created more accessible and efficient versions of these models.
A notable example is the HyperNova 60B model, now available for free on the Hugging Face platform. This model is derived from OpenAI’s gpt-oss-120B but has been compressed to approximately half its original size, reducing it to 32GB. Despite this reduction, HyperNova 60B maintains comparable accuracy and performance levels. The latest iteration, HyperNova 60B 2602, offers enhanced support for tool calling and agentic coding, areas where inference costs can be substantial.
The efficiency of HyperNova 60B is evident in its lower memory usage and reduced latency, making it a viable option for businesses seeking to implement advanced AI solutions without the prohibitive costs associated with larger models. This development is particularly significant for industries where rapid response times and cost-effectiveness are critical.
Multiverse Computing’s advancements position it as a formidable competitor in the AI sector. The company asserts that HyperNova 60B outperforms models like Mistral Large 3, developed by French AI firm Mistral AI. Both companies share similarities, including expansion beyond their home countries and a focus on enterprise clients. Multiverse Computing’s clientele includes prominent organizations such as Iberdrola, Bosch, and the Bank of Canada.
The company’s growth trajectory is underscored by reports of an impending €500 million funding round, potentially elevating its valuation to over €1.5 billion. While Multiverse Computing has confirmed ongoing discussions with potential investors, it has refrained from commenting on specific valuation or funding details at this stage. Additionally, the company has not addressed reports suggesting that its annual recurring revenue reached €100 million in January.
In comparison, OpenAI’s annual recurring revenue stands at $20 billion, while Mistral AI’s revenue has surged to over $400 million, driven by increasing demand for alternatives to U.S.-based technology solutions. Multiverse Computing positions itself as a provider of sovereign solutions across the AI stack, catering to organizations seeking independent and efficient AI capabilities.
The introduction of CompactifAI and the release of HyperNova 60B signify a pivotal moment in AI development, offering scalable and cost-effective solutions that democratize access to advanced AI technologies. As the industry continues to evolve, innovations like these are crucial in bridging the gap between cutting-edge AI capabilities and practical, affordable deployment for businesses worldwide.
