{
  "schema_version": 1,
  "currency": "USD",
  "unit": "per 1,000,000 tokens",
  "generated_from": "docs/pricing/list-prices.json@2026-09-21",
  "openai_compatible_base_url": "https://relay.obitmc.com/v1",
  "stability": "These are list prices. A change is announced with an effective date and the superseded row is kept in the source declaration, so a past invoice stays answerable. This file is generated from docs/pricing/list-prices.json in the Obit repository and is never hand-edited.",
  "models": [
    {
      "id": "unsloth/Qwen3.8-27B-GGUF",
      "name": "Qwen3.8 27B",
      "context_window_tokens": 262144,
      "max_output_tokens": 262144,
      "input_per_1m_usd": 0.1,
      "output_per_1m_usd": 1.0,
      "cached_input_per_1m_usd": 0.01,
      "effective_from": "2026-09-21",
      "notes": "Prompt and completion share one 262,144-token context window; there is no separate output cap, so max_output_tokens repeats the context window. Cached input is the price for prompt tokens served from a worker's prefix cache; the first send of a prefix pays the ordinary input price, and there is no cache-write premium."
    }
  ]
}