← run

tool-01-weather

1.000
2/2 tests· tool-calling
Challenge · difficulty 2/5
# Single tool call (weather)

Tests whether the model invokes a provided tool with correct arguments and uses the result.

The model is given a `get_weather(city)` tool and asked for the current temperature in Paris.
Scored on: (1) it calls `get_weather` with `city = "Paris"`, and (2) its final answer states
the temperature returned by the tool (18). Defined in `task.py`.
Proposed solution
TOOL CALLS:
[
 {
  "name": "get_weather",
  "arguments": {
   "city": "Paris"
  }
 }
]

FINAL:
The current temperature in Paris is 18°C, and the conditions are cloudy.