- Verses AI Inc. (CBOE:VERS) announced a major leap forward in artificial intelligence with the unveiling of its new digital brain architecture, codenamed AXIOM
- In a head-to-head showdown, AXIOM outperformed DreamerV3 across ten diverse, arcade-style environments designed to test real-world generalization with minimal data
- These results suggest that AXIOM could represent a paradigm shift in AI development—offering not only better performance but also dramatically lower computational and financial costs
- Verses AI stock (CBOE:VERS) last traded at C$5.00
Cognitive computing company Verses AI Inc. (CBOE:VERS) announced a major leap forward in artificial intelligence with the unveiling of its new digital brain architecture, codenamed AXIOM (Active eXpanding Inference with Object-centric Models). Based on the neuroscience-inspired principle of Active Inference, AXIOM has demonstrated performance that significantly surpasses Google DeepMind’s DreamerV3—one of the most advanced AI models known for its generalization capabilities in simulated environments.
In a head-to-head showdown using the Gameworld 10K benchmark—a next-generation successor to the Atari 100K Challenge—AXIOM outperformed DreamerV3 across ten diverse, arcade-style environments designed to test real-world generalization with minimal data.
AXIOM: Better, faster, cheaper (and smaller)
Verses’ AXIOM not only delivered superior gameplay performance but did so with efficiency and simplicity. Unlike traditional AI models that rely on neural networks, backpropagation, and gradient descent, AXIOM uses a novel object-centric approach rooted in Active Inference.
Performance highlights (AXIOM vs DreamerV3)
- 60 per cent better gameplay (Normalized Score: 77 vs 48)
- 7.6x more sample-efficient (3,175 steps vs 24,207)
- 39x faster GPU runtime (~10 minutes vs ~370 minutes)
- 12x cheaper to run (Estimated GPU cost: $0.66 vs $25.54)
- 400x smaller model size (0.95M vs 420M parameters)
These results suggest that AXIOM could represent a paradigm shift in AI development—offering not only better performance but also dramatically lower computational and financial costs.
Independent validation
To ensure transparency and credibility, Verses submitted AXIOM’s research paper, mathematical proofs, and source code for independent review to Soothsayer Analytics, a globally respected AI certification and advisory firm. Soothsayer is known for its rigorous standards and is trusted by Fortune 1000 and Global 2000 companies.
Expert insights
Dr. David Bray, Ph.D, chair of the Accelerator and distinguished fellow, Stimson Centre; senior fellow with the Institute for Human-Machine Cognition as well as an expert with MIT Horizon, the Oxford Internet Institute, and Harvard’s Leadership for a Networked World Program, caught early access to the paper.
“In many critical fields, there’s an urgent need for models that are more reliable, less constrained by past training sets, and more adaptable to a changing world,” He said in a news release. “Active Inference offers a more transparent, energy-efficient, and generalizable approach to intelligence—one that mirrors how humans learn, adapt, and simplify. The sample efficiency of AXIOM is particularly noteworthy, showing how it can generalize from fewer examples, much like human cognition, using far less computation. And by using Bayesian model reduction to prune complexity, this work captures a fundamental principle of intelligent behavior: simplify where possible, but not at the expense of understanding. This is the kind of intelligence we need for systems to thrive in complex, real-world environments across both public and private sectors.”
Implications for the future of AI
If AXIOM’s performance continues to hold across broader domains, it could redefine the foundations of artificial intelligence—moving away from data-hungry, compute-intensive models toward more biologically inspired, efficient systems. This could have far-reaching implications for industries ranging from robotics and autonomous systems to healthcare and smart infrastructure.
The Atari 100K Challenge, introduced in 2015, involves creating a single artificial intelligence model capable of being trained to meet or beat human-level performance on a pool of up to 26 classic Atari (OTC Pink:PONGF) video games. The AI model must learn directly from pixel data, using only the score as a reward signal. Verses’ leadership will provide broader context around benchmarking and showcase early results in comparison with other AI models.
Specializing in biologically inspired distributed intelligence, Verses AI is focused on developing next-generation artificial intelligence solutions, including the Genius platform.
Verses AI stock (CBOE:VERS) ended Monday trading more than 10 per cent higher, but also lost around 10 per cent by midday Tuesday, sitting at C$5.00.
Join the discussion: Find out what everybody’s saying about this stock on the Verses AI Bullboard, and check out the rest of Stockhouse’s stock forums and message boards.
The material provided in this article is for information only and should not be treated as investment advice. For full disclaimer information, please click here.
