Test fixture
Arithmetic, symbolic steps, and structured problem solving.
The model receives the prompt (and optional system message). The run uses scorer exact_match with the JSON configuration below. Pass/fail and partial credit are determined entirely by that scorer against the model output; no human grading.
A task starts at 09:35 and ends at 12:20 on the same day. How many minutes did it run? Return only the number.
{
"expected": "165"
}temperature
0
max_tokens
20
timeout (s)
120
type
scored
file
reasoning_medium_02.json