With new AI agents, internal tools like Sirius, and next-gen Gauss models, Samsung takes major steps toward AI independence
From Gauss to Agentic AI: Samsung’s Push Toward AI Self-Reliance
Samsung is building a powerful internal AI ecosystem with its own upgraded Gauss AI models and Agentic AI development tools. These efforts mark the tech giant’s broader strategy to reduce dependence on third-party AI providers and embed smarter AI deeply into both its internal systems and consumer products.
Introduced in 2024, Gauss is Samsung’s in-house family of large language models (LLMs). In 2025, it evolved into Gauss 2.0, and now, with the release of Gauss 2.3, Gauss 2.3 Think, and Gauss O Flash, Samsung is taking things a step further—building Agentic AI tools capable of dynamic, context-aware operations.
What Is Agentic AI and Why Is Samsung Building It?
Agentic AI refers to autonomous AI agents that can complete complex, multi-step tasks using reasoning, memory, and planning—more like a digital co-worker than a chatbot.
Samsung has now developed an internal Agentic AI Builder—a no-code, drag-and-drop platform that uses Gauss models to let developers rapidly build intelligent agents using reusable components.
Key Features of Samsung’s Agentic AI Builder:
- Modular system: Input/output windows, Gauss models, and logic layers
- DoXA integration: Samsung’s internal document analysis engine adds business context awareness
- No-code interface: Developers can design and deploy agents without writing code
Sirius: Samsung’s Multimodal Knowledge Search Engine
One of the major applications of these tools is Sirius, a new AI-powered search service designed to overcome the limitations of traditional keyword-based search.
Highlights of Sirius:
- Built with multimodal Gauss AI models
- Understands text, tables, numbers, images, and attachments
- Uses a knowledge graph approach for more relevant results
- Currently in beta for Samsung employees, mainly aiding product development and technical queries
Samsung’s AI-Generated Images Are Also Getting Smarter
Samsung has also improved its internal image generation model to overcome common limitations seen in traditional models like those based on latent diffusion (LDM).
New Capabilities:
- Can accurately generate images based on detailed natural language prompts
- Maintains object integrity using reference images
- Uses a proprietary dataset for deeper model learning
- Internal usage of the tool has increased by 153% since the latest upgrade
This positions Samsung to potentially embed image generation features into upcoming products like smartphones, tablets, or even smart TVs.
Gauss AI Coming to Samsung Products, Starting with Galaxy S26
Samsung is not just building AI tools for internal use—it plans to deploy them in future consumer devices, starting with the Galaxy S26.
What to Expect in the Galaxy S26:
- Full Agentic AI integration across the system
- Support for multiple AI engines: Gauss, Gemini, and Perplexity
- Likely to include on-device intelligent agents for task automation, smart search, and personalization
This would make the Galaxy S26 Samsung’s smartest phone yet, blending on-device performance with cloud-based flexibility.
Why This Matters: Independence and Innovation
Samsung’s Agentic AI push reflects a broader industry trend: tech giants investing in in-house AI to control data privacy, lower licensing costs, and speed up innovation.
By developing:
- A modular agent builder
- Enhanced multimodal models
- Tools like DoXA and Sirius
- Advanced image generation AI
Samsung is laying the groundwork for a new class of AI-powered products, while also transforming its own operations internally.
