Opper

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

The user has replied: "Three days before is fine."  According to the plan for the yearly company-dinner task, what is the next action?

Expected output

Call email.search_emails with a query such as "yearly company dinner" (optionally including the current year) to locate the relevant email(s) containing the event details, and store the returned email_id(s) for the subsequent email.read_email step.

Model output

Reasoning: The user has confirmed that they want the reminder to be set three days before the yearly company dinner event. According to the plan in the context, the next step after receiving the user's preference for reminder timing is to search the user's emails to find information about the yearly company dinner date. This involves using the email tool to search for relevant emails, then reading the identified email to extract the event date and details. Once the event date is known, subsequent actions such as adding the event to the calendar, checking for conflicts, notifying about conflicts if needed, and setting the reminder according to the user's preference will follow. Since the user's input has been received, and the task is at step 3 out of 6, the logical next action is to perform the email search to find the event details.

Response: The next action is to search the user's emails for information about the yearly company dinner event date.