O3 mini - Adding Reasoning Effort Parameters to AI Agents

You can achieve this now. Here's how:

1. Define the Desired Behavior:
   Determine precisely what "reasoning effort" signifies within your specific use case. For example, are you aiming to:

   • Adjust the verbosity or multi-step reasoning capabilities of a language model?
   • Modify the prompt structure to encourage a more in-depth chain-of-thought process?
   • Dynamically set internal parameters (such as temperature or maximum tokens) based on the intended reasoning level?

2. Extend Your AI Agent Node:
   If you are utilizing a custom or existing callin.io node for an AI agent (for instance, one that interacts with an AI service's API), you could adapt or encapsulate it to accept an additional parameter (e.g., reasoningEffort). Your node's code would then adjust the prompt or API parameters accordingly. For example, you might modify the prompt as follows:

```javascript
const basePrompt = "Answer the following question:";
const reasoningInstruction = reasoningEffort > 5
? " Please provide a detailed, step-by-step explanation."
: "";

const finalPrompt = ${basePrompt}${reasoningInstruction}n${userQuestion};
```

3. Generate the callin.io Workflow JSON Dynamically:
   If your application takes a natural language prompt and translates it into a callin.io workflow, you would need to incorporate logic that detects when the prompt includes a request for a specific "reasoning effort" and subsequently includes that parameter in the JSON. For instance, the JSON for an AI Agent node might appear like this:

json
{
"nodes": [
{
"parameters": {
"prompt": "Answer the following question with detailed reasoning: {{ $json.question }}",
"reasoningEffort": 7 // A custom parameter you are introducing
},
"name": "AI Agent",
"type": "callin.io-nodes-base.yourCustomAiAgent",
"position": [500, 300]
}
],
"connections": {
"AI Agent": {
"main": [
[
{
"node": "Next Node",
"type": "main",
"index": 0
}
]
]
}
}
}


In this code snippet, the node incorporates a reasoningEffort parameter that you can later leverage in your custom node logic to modify the request sent to the AI.

4. Integrate and Validate:
   After setting up your JSON generation logic, integrate it with your callin.io instance (either through the API or by importing the JSON). Ensure that the custom parameter is being correctly passed to your node and that the AI performs as anticipated.

Thanks Brian, but I believe we're discussing different aspects.

"Reasoning effort" is a parameter within the OpenAI API request that specifies the number of tokens utilized during the reasoning process.

This isn't a concept I originated; it's a new addition to the documentation.

The default setting is "medium," but o3-mini also supports "low" and "high" options. This parameter balances cost (based on the number of tokens generated) with processing speed (where "low" is significantly faster than "high").

More details can be found here:

https://platform.openai.com/docs/guides/reasoning?example=planning#advice-on-prompting

I have a very similar question. When will the OpenAI nodes be updated to support o1 and the new o3-mini? As Gregor mentioned, the new o3-mini series includes three "levels of effort": Low, Medium, and High. It would be excellent if an OpenAI node upgrade incorporated all three effort levels for the new o3-mini model.

I require this urgently as well. Kindly update the node to incorporate the latest models and enable the specification of reasoning effort.

I also require this functionality urgently. Thank you

Current version already has that