The Pulse
Giotto Launches Alto and Solo Sovereign Reasoning Models
Swiss AI lab Giotto has launched Alto and Solo, two reasoning models designed for European cloud, dedicated hosting and on-premises use. Alto runs on eight B300 GPUs, while Solo is designed for a single GPU.
AI.info Team · Written with AI · fact-checked ·
“Europe must offer more than a different flag on the same dependency. Sovereignty has to belong to the people and organisations using AI, and not just to the companies building it.”
Aldo Podestà, CEO and co-founder of Giotto
Giotto announced Alto and Solo on October 27, presenting the pair as Swiss-developed reasoning models that customers can run through European cloud services or on compatible hardware they control. The Lausanne-based AI lab says Alto is its flagship for demanding reasoning and enterprise workloads; Solo is designed to run on one GPU. The launch ties the two products to a wider pitch: let customers choose where their AI runs and where their data stays.
Two models, different hardware demands
Alto requires a server with eight B300 GPUs, according to the company’s release. Solo targets a single-GPU setup, which Giotto says makes local deployment possible on a workstation. The announcement does not specify Solo’s GPU model or memory requirement, details that would determine what hardware many prospective users need.
Giotto’s website publishes benchmark results for both models. For Solo, the company lists 96.6% on AIME 2026, 88.3% on GPQA-Diamond and 37.5% on Humanity’s Last Exam. Those are company-published figures; the website’s table does not, on the page reviewed, describe the evaluation setup in enough detail to establish an independent comparison.
Cloud access is free; private deployments are paid
Giotto says shared-cloud access is free, and its launch includes an anonymous chat service that requires no account. API use on the shared cloud carries a flat fee for unlimited requests, limited to one concurrent request; token packages are available for customers needing multiple simultaneous requests. The announcement gives no price for the flat-fee plan.
Customers seeking an isolated environment can rent dedicated cloud instances monthly, starting with one GPU, with the Giotto software licence included. On-premises installations are licensed annually or for multiple years, with pricing based on the number and type of GPUs. Giotto says deployment can take place in partner data centres in Switzerland or elsewhere in Europe, or on compatible hardware at a customer’s site.
Giotto makes control the product pitch
The company is also launching iOS and Android apps. It describes its cloud services as GDPR-aligned and says they do not depend on US cloud infrastructure. Those statements outline Giotto’s intended deployment options, but the release does not provide hosting-provider names or contract terms for customers to assess data-location guarantees.
Giotto was founded in Lausanne in 2017 by EPFL mathematics alumni, according to the release. The company also cites its results in the 2024 and 2025 ARC-AGI competitions, where it says it placed 11th and second among roughly 1,500 teams. That history supports its focus on reasoning, while the new products turn that work into a choice between a larger hosted model and a single-GPU deployment.
For buyers, the practical distinction is hardware and hosting: Alto runs on eight B300 GPUs, while Solo is intended for one GPU. Giotto has not stated the precise GPU or memory needed for Solo in its launch announcement.