# Pricing and endpoint (Chat Completions): https://opencode.ai/v2/docs/console/models/ (accessed 2026-10-21).
# Mid-conversation flags (system_updates, effort_updates, tools.updates) follow providers/xiaomi/models/mimo-v2.6-flash.toml (Zen passes requests through).
# effort_updates: left out. Changing thinking.type on the next request keeps the cache, but there's no in-history effort message, so it doesn't count.
# Tools follow Xiaomi's API: https://mimo.mi.com/static/docs/api/chat/openai-api.md (tool_choice only auto, function strict); Zen passes requests through.
schema = 2
base_model = "xiaomi/mimo-v2.6-flash"
name = "MiMo-V2.6-Flash Free"

[capabilities]
tools = { supported = true, strict = true, choice = ["auto"] }
reasoning = { supported = true }
system_updates = true

[limit]
context = 200_000
output = 32_000

[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