# Opt in with chat_template_args.enable_thinking=true. Baseten documents no
# explicit false behavior, effort values, or reasoning-token budget.
# https://docs.baseten.co/inference/model-apis/reasoning

base_model = "zhipuai/glm-5.1"
name = "GLM 5.1"

[[reasoning_options]]
type = "toggle"

[interleaved]
field = "reasoning_content"

[cost]
input = 1.3
output = 4.3
cache_read = 0.26

[limit]
context = 202_800
output = 202_800