base_model = "anthropic/claude-sonnet-4-0"
reasoning_options = []

[cost]
input = 3
output = 15
cache_read = 0.3
cache_write = 3.75

[limit]
context = 216_000
input = 128_000
output = 16_000