Tekmatix has expanded the AI models available inside AI Agent Actions, giving you more choice over how your automated AI workflows think, respond and perform.

Alongside OpenAI models, you can now select AI models from Anthropic and Google directly inside your workflow AI Agent Actions.

You can choose your preferred AI provider, select the model best suited to the task, and even control how much reasoning effort the AI applies, all from a redesigned model selection interface.

What’s New in Tekmatix AI Agent Actions?

More AI Providers and Models to Choose From

AI Agent Actions now support models from three major AI providers.

Anthropic

• Claude Opus 5

• Claude Sonnet 5

• Claude Haiku 4.5

Google

• Gemini 3.6 Flash

• Gemini 3.1 Pro Preview

OpenAI

• GPT-5.6 Luna

• GPT-5.6 Tera

• GPT-5.6 Sol

• GPT-5 Nano

This means you are no longer limited to choosing between models from a single AI provider.

Instead, you can select the model that makes the most sense for the specific job your AI Agent needs to perform.

A Redesigned AI Model Picker

The model selection interface inside Tekmatix has also been redesigned to make choosing the right AI model much easier.

Models are now grouped by provider, allowing you to quickly compare options from Anthropic, Google and OpenAI.

You’ll also see additional information to help guide your decision.

Models that support advanced reasoning are identified with a Thinking indicator, while short descriptions explain what each model is best suited for.

This makes it easier to choose between models designed for deeper reasoning, speed, everyday automation tasks or more complex AI operations.

Control How Much Reasoning Your AI Uses

For supported models, you can now choose the amount of reasoning effort the AI should apply.

Available reasoning levels include:

• Low

• Medium

• High

Higher reasoning effort allows the AI to spend more computational effort working through complicated problems and multi-step tasks.

Lower reasoning effort can be useful when you need fast responses for straightforward activities.

This gives you another level of control over how individual AI Agent Actions operate inside your Tekmatix workflows.

For example, you might use a higher reasoning level for:

• Complex decision-making

• Detailed data analysis

• Multi-step problem solving

• Evaluating several pieces of information before taking an action

• Tasks where accuracy is particularly important

A lower reasoning level may be better suited to:

• Intent classification

• Basic information extraction

• Simple data processing

• Straightforward CRM actions

• Routine workflow operations

  

How to Choose an AI Model in Tekmatix

To change the model used by an AI Agent Action:

1. Open the relevant Workflow inside Tekmatix.

2. Open your AI Agent Action.

3. Click the Model dropdown.

4. Browse the available models grouped by AI provider.

5. Review the model description and Thinking indicator where available.

6. Select the AI model you want to use.

7. Choose a Low, Medium or High reasoning effort level if the selected model supports reasoning controls.

8. Save your changes.

9. Publish your updated workflow.

Your AI Agent Action will then use the model and reasoning settings you selected whenever that workflow step runs.

Which AI Model Should You Use?

There isn’t necessarily one "best" AI model for every Tekmatix workflow.

The better approach is to match the model to the job.

A powerful flagship model may be appropriate when your AI Agent needs to analyse complex information, follow multiple instructions or make decisions where reasoning quality is particularly important.

A faster, lighter model may be a better choice for high-volume tasks such as categorising leads, extracting simple information, identifying intent or performing routine operations.

This allows you to balance:

• AI capability

• Reasoning depth

• Response speed

• Token consumption

• Workflow complexity

• Cost

You can make these decisions independently for each AI Agent Action rather than using the same model across every automation.

Why This Matters for Tekmatix Users

Giving businesses access to multiple AI providers makes AI automation much more flexible.

Different models have different strengths.

You may find one model performs particularly well for writing and communication, while another is better suited to reasoning, classification, data analysis or high-speed processing.

With Anthropic, Google and OpenAI models now available inside Tekmatix AI Agent Actions, you can test different approaches and build an AI workflow stack that matches the way your business actually operates.

The addition of reasoning effort controls gives you even more flexibility.

Instead of simply choosing a model, you can also decide how much computational effort that model should apply to each individual task.

For businesses running large numbers of AI-powered workflows, this can help you avoid using maximum reasoning power for simple jobs that simply don’t require it.

Combine Smarter Model Selection With Lower AI Agent Costs

This update builds on the recent optimisation of Tekmatix AI Agent Actions, which reduced the cost of running AI Agents by up to 50% in internal testing, depending on the task and configuration.

Improvements such as prompt caching, leaner system instructions and more efficient model usage have helped reduce unnecessary token consumption.

Now, with additional providers and reasoning controls available, you have even more ability to choose the right balance between intelligence, speed and cost for each automation.

That could mean using a powerful reasoning model for a critical decision-making workflow while assigning faster and lighter models to repetitive tasks running hundreds or thousands of times.

Build Smarter AI Workflows With Tekmatix

AI automation isn’t just about adding AI to every workflow.

It’s about choosing the right AI for the right job.

With Anthropic, Google and OpenAI models available inside AI Agent Actions, Tekmatix gives you greater control over which models power your automations, how deeply they reason and how efficiently they operate.

If you already use AI Agent Actions, open one of your existing workflows and explore the new model picker to see which models and reasoning settings best suit your automations.

If you haven’t started experimenting with AI Agent Actions yet, this is a great opportunity to begin building workflows that can analyse information, make decisions, generate responses and take action throughout your business automatically.