--- title: "Intility Inference | Intility" description: "Gain access to leading open AI models through a secure, fully managed inference service, provided by Intility." locale: en url: https://intility.com/en/what-we-do/artificial-intelligence/inference updated: 2026-10-21T13:45:35.402Z --- # Intility Inference Gain access to leading AI models through a secure, fully managed inference service, provided by Intility from Norwegian data centers under Norwegian jurisdiction. ## Select model. Get a private endpoint. **Model Catalogue** Access leading open models through OpenAI-compatible APIs. Intility develops, secures, and operates the entire service – from underlying infrastructure and model management to monitoring, support, and continuous development. Choose a model tailored to the specific task. ### Qwen3 Embedding 8b **32k context** **Embedding** A multilingual embedding model for semantic search, information retrieval, classification, and utilizing the organization’s own data in AI solutions. **Price per 1m tokens** - **Input**: NOK 1,00 - **Cached Input**: NOK 0,30 - **Output**: n/a ### GLM-5.3 Flash **400k context** **Multimodal input** A cost-efficient, natively multimodal model for reasoning, coding, and agent-based workflows. Combines text and image understanding with strong tool-use capabilities for software development, visual analysis, and complex tasks **Price per 1m tokens** - **Input**: NOK 10,00 - **Cached input**: NOK 2,00 - **Output**: NOK 40,00 ### GLM-5.3 **400k context** **Cyber Capabilities** Z.AI's next generation open top model, developed for advanced reasoning, software development, agent-based tasks, and cyber defense. The model combines strong code understanding with the ability to identify vulnerabilities, use tools, and solve complex tasks over time. **Price per 1m tokens** - **Input**: NOK 20,00 - **Cached input**: NOK 5,00 - **Output**: NOK 75,00 ## Use Cases for Intility Inference Connect leading open models to the organization's own applications, data, and workflows. Build new AI features, analyze and structure large amounts of information, make the organization's knowledge more accessible, and automate tasks that previously required manual processing. ### Integrate AI into your own applications Incorporate generative AI into new and existing applications, products, specialized systems, and digital services. ### Analysis of own data Use language models to analyze, summarize, classify, and structure large amounts of text and other information. ### Search and knowledge retrieval Combine embedding models and language models to find and utilize relevant information from the organization's own data sources. ### Automate workflows Use AI to process information, classify inquiries, and automate repetitive tasks in established workflows. ### Develop software with AI Connect OpenCode and other developer tools to Intility Inference for code generation, debugging, and complex development tasks. ### Use AI agents Develop agents that can reason, use tools, and perform tasks across the organization's applications and data sources. ## Enterprise AI requires an enterprise platform **Security & Compliance** Intility Inference is delivered as an integrated part of an established enterprise platform, developed, secured, and operated by Intility. The platform is already used by hundreds of businesses, including critical and regulated entities with high demands for security, privacy, compliance, availability, and operational control. From the underlying infrastructure and model management to monitoring, support, and continuous development, the entire service is delivered as a single cohesive offering. ### Norwegian data centers Models and customer data are processed in secure, Norwegian data centers with high operational reliability and control. ### Norwegian jurisdiction The service is provided by a Norwegian supplier and is fully subject to Norwegian law and jurisdiction. ### Integrated security Security is built into the entire delivery, from the underlying infrastructure to ongoing monitoring and incident management. ### Ready for production Transition from experimentation to mission-critical workloads without switching platforms or building a new underlying architecture. ## Documented security and compliance built into the service Intility Inference is developed and operated within Intility's established framework for information security, risk management, and operational control. The service follows documented processes for access control, change management, monitoring, incident management, and continuous improvement. This provides the organization with a clear and documentable security and control environment around the use of open AI models – from underlying infrastructure and data processing to ongoing operation and further development of the service. - **ISAE 3402 Type II**: Independent attestation of controls related to information security - **ISAE 3000 Type I**: Independent attestation of established controls and processes - **ISAE 3000 Type II**: Independent attestation of controls related to privacy and data processing - **Trusted Cloud Provider**: Registered in the Cloud Security Alliance STAR Registry - **NSM Basic Principles**: The platform is developed and operated in accordance with NSM's fundamental principles for ICT security - **ISO 27001**: Certified information security management system ## Change the URL, don't rebuild the solution The API is compatible with the OpenAI Chat Completions API. Point to an SDK or framework you are already using against the endpoint at Intility, and keep the rest of the application architecture. ### /v1/chat/completions The widest compatibility across SDKs, frameworks, and tools. ### /v1/responses For OpenAI SDKs configured against the Responses API. ### Streaming Receive tokens continuously as server-sent events, so that the responses can be displayed as they are generated. ### Tool calling Define tools and let the model return structured function calls that the application can execute. ### JSON mode Limit the model's responses to structured and machine-readable JSON for use in applications and automated workflows. ### Thinking mode Enable extended reasoning for complex tasks and receive a structured summary of the reasoning along with the answer. ## The people behind the platform **BUILT AND OPERATED BY INTILITY** Behind Intility Inference is a dedicated professional environment responsible for the entire delivery – from security, platform operations, and model management to consulting and integration with the business's own solutions. ### 24/7 Security Operations Center Intility's Security Operations Center monitors the inference platform and handles security incidents around the clock. ### Platform Engineers Dedicated platform specialists operate, maintain, and further develop the infrastructure, models, and associated services. ### AI Specialists People working with these models daily, so a model change or end of life reaches you as advance notice rather than a surprise. ### Consultants Strategic advisory and hands-on expertise, so the platform turns into working solutions inside your own systems. **FAQ** ### Frequently Asked Questions Here you will find answers to some of the most common questions about models, data, integration, pricing, and access. Please contact us if you have any further questions. ### What is inference? Inference is the process of using a trained AI model to process new data and generate an output. For example, when you send a prompt to a language model and receive a response, the model is performing AI inference. Simply put, training is when an AI model learns, while inference is when it uses what it has learned. AI inference can be used for tasks such as text generation, image analysis, classification, embeddings, and other AI applications. ### Is Intility Inference compatible with OpenAI? Yes. Intility Inference offers OpenAI-compatible APIs and can be used with established SDKs, frameworks, and development tools. Existing integrations can connect to the service by changing the endpoint and API key. ### Where is the data processed? Models and customer data are processed in secure data centers in Norway. The service is provided by Intility under Norwegian jurisdiction. ### Is the data used to train the models? No. Data processed through Intility Inference is not used to train the underlying models. ### What models are available? Available models and pricing are listed in the model catalog. New models are continuously made available as technology and needs evolve. ### What happens when a model is phased out? Phasing out of models is announced in advance and managed as part of Intility's model governance. This allows time to test and adopt an alternative model through the same integrations. ### What does Intility Inference cost? Usage is priced per million tokens. Updated prices for input, cached input, and output are listed in the model catalog. ### How do I gain access to Intility Inference? Contact Intility to set up access and a private endpoint. ## Contact us for more information about Intility Inference ### Ulrik Wilhelm Koren - [Tel: +47 93 63 08 87](tel:+4793630887) - [Mail: ulrik.wilhelm.koren@intility.no](mailto:ulrik.wilhelm.koren@intility.no)