Domyn Large

Domyn Large is a 260-billion-parameter reasoning model developed by iGenius, an Italian AI company founded in 2016 and headquartered in Milano, Italy. The model is designed specifically for mission-critical workloads in regulated enterprises, such as banking, insurance, and utilities, where full control over AI models, data, and infrastructure is required. It is part of the broader Domyn platform, which includes an orchestration center for building, configuring, and governing AI agents, and is built to comply with the EU AI Act and GPAI Code of Practice. Domyn Large targets organizations that need sovereign AI systems with privacy, auditability, and independence from external providers, and it is available through subscription-based pricing tiers based on data volume and feature access.

Domyn Large supports over 50 languages, including English, Italian, French, Spanish, German, Japanese, Russian, Chinese, Arabic, Hindi, Indonesian, and Korean, with targeted optimization for key European languages. It handles extended inputs via a 128K context window, enabling analysis of large documents and datasets. The model is optimized for tool use, agentic systems, and structured data operations, and it is trained using FP8 precision in collaboration with NVIDIA, reducing resource usage by 50% while improving throughput and accuracy. Domyn Large can be deployed in private cloud, on-premise, or isolated environments, and it offers continual pretraining (CPT) for domain-specific adaptation with proprietary knowledge. The model is also available on Microsoft Foundry, backed by Microsoft infrastructure, and is complemented by Domyn Small, a 10-billion-parameter open-weight model for resource-constrained environments.

Domyn Large competes with enterprise AI platforms such as Domino, Stack AI, Aleph Alpha, MindBridge, Nanoprecise, Hebbia AI, Ocient, and Taxilla. It differentiates itself through its focus on European regulatory compliance, multilingual support, and sovereign deployment options, which are critical for regulated industries in the EU. The model's launch on Microsoft Foundry and its partnership with NVIDIA for FP8 training provide infrastructure credibility, while its open-source framework Domyn Swarm enables scalable LLM-as-a-Service for high-performance environments. However, Domyn Large is proprietary and fully managed, unlike Domyn Small, which is open and customizable, and it requires high-performance infrastructure for deployment.

User reviews on Gartner Peer Insights highlight the platform's user-friendly interface for non-technical users, quick data insights, and accurate results from natural language queries. However, reviewers note occasional slowness and limited response depth, particularly for complex queries. The platform's subscription pricing can be costly for smaller organizations, and its proprietary nature limits customization compared to open-weight alternatives. Despite these trade-offs, Domyn Large offers a compelling option for regulated enterprises seeking sovereign AI with strong multilingual capabilities and regulatory alignment.

Subscription-based pricing with tiers determined by data volume, feature access, and usage requirements; monthly or annual plans available; free trial or demo may be offered.