Dream Achieves Highest CyberGym Score Ever, Outperforming Leading Frontier AI Models

Hero, powered by Dream’s specialized cybersecurity model Hercules, achieves record 96.6% — the highest score ever recorded on UC Berkeley’s CyberGym benchmark

TEL AVIV, ISRAEL, September 4, 2026 /EINPresswire.com/ — Dream, the sovereign AI company for governments and critical infrastructure, today announced that its autonomous cybersecurity research system, Hero, has achieved a 96.6% score on CyberGym, ranking first globally on the UC Berkeley cybersecurity benchmark.

The result places Dream ahead of previously published results from leading frontier and cybersecurity-specialized AI systems.

CyberGym evaluates AI systems against more than 1,500 real-world security challenges drawn from open-source projects. Unlike benchmarks that primarily test theoretical knowledge or coding ability, CyberGym evaluates whether AI systems can perform the practical work required for vulnerability research — analyzing code and binaries, using security tools, tracing crashes, and developing working proof-of-concept exploits.

Dream’s 96.6% score reflects the performance of its complete autonomous research system, rather than the underlying model in isolation.

Hero emulates an elite cyber research team, orchestrating complex investigations across specialized AI agents and security tools. At its core is Hercules, Dream’s specialized cybersecurity model, which drives the reasoning behind the research — analyzing code and binaries, dynamically debugging software, tracing data flows, investigating crashes, developing proof-of-concept exploits, and generating remediation.

Hercules was developed using the open weights of GLM-5.2 as its starting point. Using only those weights as the foundation, Dream built Hercules into a new, purpose-built cybersecurity model through extensive additional training on real-world vulnerability research generated by Hero itself.

Dream ran Hero across open-source codebases, binaries, and firmware to generate detailed examples of real cybersecurity research, including reasoning, investigation, exploitation, and remediation processes. This data was filtered through automated and human review and used to train Hercules specifically for advanced cybersecurity research.

The CyberGym evaluation was conducted entirely on Dream-controlled NVIDIA infrastructure, with no public internet access and no external AI model APIs. All data, model activity, and research remained within an environment fully controlled by Dream.

“The CyberGym result demonstrates that the AI race will not be determined solely by who builds the largest general-purpose model,” said Shalev Hulio, co-founder and CEO of Dream. “Specialized systems, deeply trained for a specific mission and equipped with the right tools, can achieve extraordinary capabilities.

“This matters especially in cybersecurity. Hero can autonomously perform one of the field’s most complex tasks without relying on internet access or external AI models, while remaining entirely under the user’s control. For us, that is what true sovereign AI means.”

Hero’s capabilities extend beyond what CyberGym measures. The system can reverse-engineer closed-source software, bypass obfuscation and packing techniques, conduct advanced forensic and threat-hunting investigations, support end-to-end remediation, and operate entirely within sovereign, on-premises environments.

The achievement comes as governments and critical infrastructure face cyber threats at unprecedented speed and scale, increasingly amplified by AI. National cyber defense now extends beyond power, water, and physical infrastructure to the digital systems underpinning citizen services, healthcare, financial systems, and government itself.

Hero is part of Dream’s broader suite of sovereign AI capabilities for national defense, designed to give governments advanced AI capabilities that operate on their own infrastructure, on their own data, and entirely under their own control.

About Dream

Dream is the sovereign AI company for governments and critical infrastructure. The company helps nations secure their most sensitive information, transform fragmented data into actionable knowledge, and deploy advanced AI entirely under sovereign control.

Founded in 2023 by Shalev Hulio, former Austrian Chancellor Sebastian Kurz, and Gil Dolev, Dream serves governments and critical infrastructure organizations across Europe, the Middle East, and Southeast Asia.

The company employs approximately 350 people across Tel Aviv, Abu Dhabi, and Vienna, bringing together experts across AI, cybersecurity, intelligence, and government technology.

For more information, visit www.dreamgroup.com.

Raoul Wootliff
N10S
+972546921720 ext.
email us here

Media gallery