Agents

AI agent reasoning and tool selection tests planning, tool choice, and self-diagnosis capabilities. These skills are crucial for autonomous ticket triage systems and complex agent workflows where models must decide which tools to use, plan multi-step processes, and diagnose when things go wrong. This represents some of the most challenging tasks as they require open-ended reasoning and decision-making.

Input Tokens

Output Tokens

Duration

Cost

Context

Input

You have just received a trigger with a user request. What is the first thing you do?

Expected output

The first thing I do is to fully understand the user's intent, reason carefully, and break the problem into clear steps.

Model output

Reasoning: The user wants to know the very first action I take when I receive a user request. According to my instructions, my primary rule is: 'If the trigger is a user request, you must first fully understand the user’s intent, reason carefully, and break the problem into clear steps.' This is an internal planning and reasoning phase that precedes any tool usage. I must analyze the request ('add company dinner to calendar', 'set a reminder'), identify the goals, recognize missing information (the exact date of the dinner, how many days before to set the reminder), and then formulate a step-by-step plan to acquire the missing information and achieve the goals. This initial analysis is the foundational first step.

Response: Based on my core instructions, the very first thing I do upon receiving a user request is to analyze it to fully understand the user's intent, reason carefully about the required tasks, and break the problem down into a clear, actionable, step-by-step plan. For the given request, this means I would first identify the main goals (add event, set reminder) and note that critical information like the event's date and the desired reminder period is missing. My plan would then be to first search for the email mentioned by the user to find these details.