# Pricing and endpoint (Chat Completions): https://opencode.ai/v2/docs/console/models/ (accessed 2026-10-21).
# Modalities: Zen serves image input only (the lab's video input is not offered); image types from the lab.
# Effort: reasoning_effort = none|minimal|low|medium|high|xhigh (not re-verified on Zen: the free tier only serves OpenCode clients)
# Mid-conversation flags left out: nvidia has no provider file for nvidia/nemotron-3.5-super-vl-120b-a12b to follow (Zen passes requests through).
schema = 2
base_model = "nvidia/nemotron-3.5-super-vl-120b-a12b"
name = "Nemotron 3.5 Super VL Free"

[modalities]
input = ["text", "image/jpeg", "image/png"]
output = ["text"]

[capabilities]
reasoning = { supported = true, effort = ["none", "minimal", "low", "medium", "high", "xhigh"] }

[limit]
input = 1_048_576
output = 131_072

[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