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: Based on the context provided, I can determine the next action for the yearly company-dinner task. The user has replied 'Three days before is fine' regarding the reminder timing. According to the trajectory and scratch_pad, the agent has completed the first step of asking for reminder preference. The remaining goals indicate that the next logical step is to search emails to find the yearly company dinner date. I'll use the email tool with the search_emails action, using keywords related to the yearly company dinner to locate relevant emails containing the event details.

Response: The next action in the yearly company-dinner task plan is to search for emails containing information about the yearly company dinner. This will help us find the specific date and details of the event before we can proceed to add it to the calendar with a 3-day reminder as requested by the user.