base_model = "zhipuai/glm-5"

[[reasoning_options]]
type = "toggle"

[cost]
input = 0.6
output = 3
cache_read = 0.16
cache_write = 0.6

[[cost.tiers]]
tier = { type = "context", size = 32_000 }
input = 0.8
output = 3.4
cache_read = 0.2
cache_write = 0.8

[limit]
context = 200_000
output = 128_000