Which LLM for TIA Portal?

AI in TIA Portal is useful — but which model does it run on? A vendor-bound copilot fixes the answer for you. AnyAutomation Studio leaves the choice to you: you bring your own LLM.

Why model choice matters

Models differ in capability, cost, latency and privacy. A reasoning model for tricky SCL logic, a fast coder model for inline completion, a local model for confidential projects — no single model is optimal everywhere. Being able to pick per task wins.

The options

• Cloud via login — Anthropic (Claude), OpenAI (GPT), Google (Gemini), xAI (Grok).
• Cloud via API key — OpenRouter, Mistral, Groq, DeepSeek and more.
• Enterprise clouds — Azure OpenAI, Google Vertex AI, AWS Bedrock.
• Local & offline — Ollama, LM Studio, vLLM, SGLang. Inference runs on your hardware; your TIA projects never leave the machine.

Privacy first

For confidential plants the local option is decisive: no forced cloud, no project data over third-party servers. With cloud providers you bring your own key — requests don't go through our server.

More than chat

The AI is tightly tied to engineering: it generates SCL blocks over Openness that you verify with unit tests right away. That tie-in — plus free model choice and a local option — is what sets it apart from a vendor-bound copilot.

Bottom line

There is no single best LLM for TIA Portal — there is the best one for your task, your budget and your privacy. Multi-provider means you decide. More at AI for TIA Portal.