HyperNova 60B 2602, a 50% compressed version of OpenAI's gpt-oss-120B, accelerates Multiverse's plans to deliver hyper-efficient, high-performance models for free to developers
DONOSTIA, Spain, Feb. 24, 2026 (GLOBE NEWSWIRE) -- Multiverse Computing, the leader in AI model compression, today announced the release of HyperNova 60B 2602 for free on Hugging Face. Building on the company's original HyperNova 60B released in January, the model is a 50% compressed version of OpenAI's gpt-oss-120B and now delivers improvements in tool calling and agentic coding compared to the previous version. The release marks Multiverse's continued commitment to making high-performance AI accessible to the global developer community, with additional compressed models planned for release in 2026.
Developers are increasingly constrained by the infrastructure requirements of massive LLMs, limiting their ability to test, refine, and deploy AI in meaningful ways. Multiverse focuses on building efficient, high-performance models that reduce these barriers while preserving advanced reasoning capabilities, releasing compressed versions of the world's most powerful models for free and eliminating compromises and trade-offs between size, accuracy, and accessibility.
At the core of this approach is Multiverse's proprietary compression technology, CompactifAI, which applies quantum-inspired mathematics to analyze and reorganize neural networks, preserving only their most information-rich components. It is capable of reducing model size by up to 95% while maintaining precision within a 2-3% margin, a stark contrast from the industry standard of 20-30% accuracy loss with 50-60% compression techniques. This enables models to retain intelligence while operating with a fraction of the compute, memory, and power typically required.
"The launch of HyperNova 60B 2602 demonstrates compression as an iterative process of improvement, not a one-time optimization," said Enrique Lizaso Olmos, CEO of Multiverse Computing. "Each generation of our compressed models pushes the boundaries of what's possible with efficient AI. By continuing to refine HyperNova and releasing it openly, we're giving developers the tools to experiment, validate, and deploy efficient AI without massive infrastructure investments."
Following strong global response to HyperNova 60B, which demonstrated that frontier-level reasoning performance could be achieved with dramatically lower memory, energy, and infrastructure requirements at a fraction of the typical cost, HyperNova 60B 2602 delivers enhanced performance and deployability, reflecting additional optimization informed by real-world developer usage and feedback.
HyperNova 60B 2602 features substantial improvements in key benchmarks compared to the previous version for tool calling capabilities, enabling more advanced agentic workflows and coding applications, including:
- Tau2-Bench: 5x improved agentic tool use
- Terminal Bench Hard: 2x improved agentic coding & terminal use
- BFCL v4: 1.5x improved function calling
This combination of intelligence and efficiency at scale validates the viability of compression for production AI deployment. HyperNova 60B 2602 retains nearly the same tool-calling capability as the full-size OpenAI gpt-oss-120B while operating at half the size, from 61GB to 32GB.
With this latest model launch, Multiverse continues to expand access to production-ready AI models designed for real-world deployment across enterprise, research, and public sector environments. Multiverse also plans to continue releasing additional open-source models and updates throughout the year, extending its approach across multiple sizes and use cases, from enterprise systems to edge and device-level applications.
The company is uniquely positioned to deliver sovereign solutions across the AI stack, and their open-source model release strategy is also designed to support the growing global community of developers, IT leaders, and professionals evaluating AI for internal and commercial use. Open access allows organizations to assess performance, security, and operational fit before committing to large-scale implementation, with minimal integration effort required, enabling greater control and independence.
For developers and professionals interested in testing Multiverse Computing's models, all releases will be available on Hugging Face at https://huggingface.co/MultiverseComputingCAI. HyperNova 60B 2602 can be accessed now at https://huggingface.co/MultiverseComputingCAI/Hypernova-60B-2602.
Technical documentation, benchmarks, and integration guides accompany each release on the company's Hugging Face page, https://huggingface.co/MultiverseComputingCAI. To learn more about compressed AI models, visit multiversecomputing.com.
About Multiverse Computing
Multiverse Computing is the leader in quantum-inspired AI model compression. The company's deep expertise in quantum software led to the development of CompactifAI, a revolutionary compressor that reduces computing requirements and unleashes new use cases for AI across industries. Headquartered in Donostia, Spain, with offices in the United States, Canada, and across Europe, Multiverse serves more than 100 global customers, including Iberdrola, Bosch, and the Bank of Canada. For more information, visit www.multiversecomputing.com.
Media Contact
LaunchSquad for Multiverse Computing
multiverse@launchsquad.com
Photos accompanying this announcement are available at:
https://www.globenewswire.com/NewsRoom/AttachmentNg/d7edb645-cb34-4dab-a0fc-d1d0f593364c
https://www.globenewswire.com/NewsRoom/AttachmentNg/70fa9621-4aac-4051-afb0-099cf4a15529
