What does Giotto cost?

Freemium · from €2.10 per million input tokens. On 29 Oct 2026, https://www.giotto.ai/api listed Giotto Alto, with EUR selected, at €2.10 per million input tokens, €10.50 per million output tokens, €0.21 per million cache-hit tokens, and €2.10 per million cache-write tokens. The page also has USD and CHF buttons. The October 27, 2026 press release says shared cloud access is free, that Alto is free to everyone, and that a free anonymous chat needs no account. It describes shared-cloud API access as a flat fee for one concurrent all-you-can-use request, without a euro amount. https://www.giotto.ai/offering does not print a dedicated-cloud or on-premises price.

Chat assistants Local models

Giotto

4.2/ 5

Lausanne lab's Alto and Solo reasoning models, with free shared cloud and on-premises options.

Also known as Giotto Alto, Giotto Solo, Giotto.ai

What it does

Giotto is a Lausanne AI lab's reasoning product. Its October 27, 2026 press release announces Giotto Alto and Giotto Solo, Swiss-developed reasoning models for shared cloud, dedicated cloud, and on-premises use. The release says Alto, the flagship for demanding reasoning and enterprise workloads, runs on eight B300 GPUs and is available for free to everyone. It says Solo is designed to run on a single GPU, and calls Solo the smartest model to run on one GPU. Shared cloud access is free. A free anonymous chat requires no account, and Giotto is launching apps for iOS and Android. The release says shared-cloud API access uses a flat fee for all-you-can-use requests, limited to one concurrent request, with token packages when more than one request runs at a time, and it does not print a euro amount for that fee. Dedicated cloud instances start at one GPU, are rented monthly, and include the software licence. On-premises licences are annual or multi-year and are priced by the number and type of GPUs. The giotto.ai homepage publishes Giotto's own benchmark tables. For Alto it lists AIME 2026 at 100.0%, GPQA-Diamond at 91.9%, HLE at 52.0%, and Math-500 at 99.80%. For Solo it lists AIME 2025 at 100.0%, AIME 2026 at 96.6%, GPQA-Diamond at 88.3%, HLE at 37.5%, and Math-500 at 99.80%. The API page, with EUR selected, lists Giotto Alto at €2.10 per million input tokens, €10.50 per million output tokens, €0.21 per million cache-hit tokens, and €2.10 per million cache-write tokens. Its sample calls the model giotto-model. That page also has USD and CHF buttons. The offering page describes dedicated European cloud, a download for private GPUs, and Giotto-ready hardware, and does not list a price. Neither the press release nor those pages list a parameter count or a context window.