Fortytwo research lab today announced benchmarking results for its new AI architecture, known as Swarm Inference. Across key AI evaluation tests — including GPQA Diamond, MATH-500, AIME 2024, and LiveCodeBench — the Fortytwo networked model outperformed OpenAI’s ChatGPT 5, Google Gemini 2.5 Pro, Anthropic Claude Opus 4.1, xAI Grok 4 and DeepSeek R1, demonstrating stronger reasoning abilities even under conditions where other frontier models fail.
Swarm Inference operates through a network of interconnected models, answering as one. The network currently consists of hundreds of nodes, run by a community of AI enthusiasts worldwide. Each node hosts a small language model (SLM) selected by its operator. The SLM can be fully custom-built specialized model, a fine-tuned version of an existing model, or a publicly available open-source model.
When a prompt is introduced, multiple nodes respond, their outputs are ranked by each other, and the highest quality answers are combined. To the outside observer, the network behaves like a single model, though in reality it emerges from the coordination of hundreds of independent SLMs.
To ensure the model’s accuracy under varied conditions, Fortytwo also conducted additional benchmark tests that included extraneous context alongside standard benchmark prompts.
This type of testing is similar to methods used in university exams and olympiads and is known as ‘extraneous information’ problems where the conditions include additional irrelevant information that is not required for solving the task. This helps determine whether the student genuinely understands the essence of the problem or is simply applying familiar formulas mechanically.
In this scenario, frontier AI models showed steep declines in accuracy, often getting trapped in repetitive reasoning loops or misled by irrelevant details. In contrast, Fortytwo’s Swarm Inference maintained stable accuracy, demonstrating a better alternative to individual LLM reasoning. By coordinating peer-ranked responses across diverse models, Swarm Inference enables intelligence to scale beyond the current limits of reasoning approaches and opens a new path for reliable problem-solving.
Fortytwo’s Swarm Inference demonstrates that intelligence can arise from a decentralized network of small, diverse models, coordinated by peer validation, rather than from ever-larger centralized systems.
"These results represent more than just outperforming frontier models; they mark the emergence of a new approach to building AI,” said Ivan Nikitin, co-founder and CEO of Fortytwo. “This breakthrough paves the way for scalable, reliable, and community-driven artificial intelligence — establishing a foundation for the next era of AI.”
Fortytwo will continue to grow the network, enabling fully open participation by node operators, custom model providers, and data scientists. The team is pushing for even greater accuracy and intelligence across the swarm, with an API release planned later this year to rival frontier AI companies in the most demanding use cases: coding, deep research, and advanced reasoning.
About Fortytwo
Fortytwo is a decentralized AI network that introduces a new type of AI inference, called Swarm Inference, where small language models running on consumer-grade hardware collaborate to achieve scale and reasoning capabilities beyond leading AI models. The entire network acts as a single model, becoming more powerful and efficient with each new node, while operating in a permissionless, transparent, and censorship-resistant manner. Anyone can contribute by running a node, with node performance ratings recorded on the Monad blockchain for evaluation and security. Developers can access frontier AI reasoning through API or on-chain, while node operators are rewarded for their participation. For more information, visit fortytwo.network.
