Logical Intelligence, an artificial intelligence company focused on energy-based reasoning systems, has unveiled Kona 1.0, its first energy-based model (EBM) designed for advanced reasoning. The model will enter pilot programs with select partners in the energy, advanced manufacturing, and semiconductor sectors later this quarter. A live demo of Kona 1.0 is now available on the company’s website, starting with head-to-head Sudoku challenges against leading large language models, with future demonstrations planned for chess and Go. Eve Bodnia, founder and CEO of Logical Intelligence, emphasized that Kona 1.0 learns by identifying and correcting its own errors rather than relying on probabilistic guessing. “If general intelligence means the ability to reason across domains, learn from error, and improve without being retrained for each task, then we are seeing in Kona the first credible signs of AGI,” she said. She added that while Kona is not the final form of AGI, it marks a significant departure from narrow AI and represents a foundational step toward a broader ecosystem of complementary AI systems. The company has strengthened its leadership with two key appointments. Yann LeCun, former chief AI scientist at Meta and recipient of the 2018 ACM Turing Award, has joined as founding chair of the Technical Research Board. Patrick Hillmann, formerly Chief Strategy Officer at Binance and an executive at General Electric and Edelman, has been named Chief Strategy Officer. LeCun, a pioneer in AI research, stated that true reasoning should be framed as an optimization problem—exactly what energy-based models aim to achieve by minimizing an energy function. “Logical Intelligence is the first company to move EBM-based reasoning from theory to real-world products, paving the way for more reliable and trustworthy AI systems,” he said. Hillmann highlighted the growing need for AI systems that are not only high-performing but also certifiable, auditable, and safe. “AI is entering domains where failure has real-world consequences,” he said. “We’re engaging early with policymakers and industry leaders to ensure responsible deployment and scalable governance.” Kona 1.0 builds on Logical Intelligence’s existing work in formal verification and verified code generation, which has already been used to prove software correctness in high-stakes, regulated environments. Unlike traditional models that predict the most likely output, Kona 1.0 maps constraints and boundaries, then finds solutions that satisfy all conditions—enabling provable correctness even as systems evolve. The company’s team also includes Fields Medalist Michael Freedman as Chief of Mathematics and Vlad Isenbaev, ICPC World Champion and former Facebook and Nuro engineer, as Chief of AI. Logical Intelligence develops AI for mission-critical applications where reliability is non-negotiable—such as energy infrastructure, semiconductor design, robotics, and industrial automation. Its Aleph agent is already available for formal verification and automated code generation with machine-checkable proofs. Kona 1.0 extends this approach into large-scale, continuously verifiable reasoning systems.