# Endpoint (Chat Completions) and free pricing match the former LL Alpha Free listing.
# Effort: reasoning_effort = low|medium|high|xhigh|max (not re-verified on Zen: the free tier only serves OpenCode clients)
# Mid-conversation flags left out: recursive has no provider file for recursive/flint to follow (Zen passes requests through).
schema = 2
base_model = "recursive/flint"
name = "Flint 40B Free"

[capabilities]
reasoning = { supported = true, effort = ["low", "medium", "high", "xhigh", "max"] }
structured_output = true

[limit]
output = 65_536

[api.chat-completions]
base_url = "https://opencode.ai/inference/openai/v1"
compatibility = { reasoning_field = "reasoning_content" }

[cost.text.input]
price = 0
cache = { read = 0 }

[cost.text.output]
price = 0