base_model = "google/gemini-2.5-flash"
# Chat uses `thinking.budget_tokens` 0..24576; 0 disables and -1 requests dynamic
# thinking in native Gemini, but -1 is a sentinel rather than a numeric budget.
# https://ai.google.dev/gemini-api/docs/thinking (accessed 2026-06-25)
reasoning_options = [{ type = "budget_tokens", min = 0, max = 24_576 }]

[cost]
input = 0.3
output = 2.5
cache_read = 0.03
input_audio = 1