Keeping the chain of thought alive across turns

Most chat apps throw away the model's thinking after each turn. Ox Alpha lets you keep it: when you send the next message, include the previous assistant turn with its reasoning_details and the model picks up its own train of thought.

The pattern

[
  { "role": "user", "content": "How many r's in 'strawberry'?" },
  {
    "role": "assistant",
    "content": "There are 3.",
    "reasoning_details": [ /* pass back unmodified */ ]
  },
  { "role": "user", "content": "Are you sure? Think again." }
]

Send that with "reasoning": { "enabled": true } and the second answer builds on the first round of thinking instead of starting cold.

Why it matters

For multi-step debugging or agent loops, continuity means the model doesn't re-derive context it already worked out. It's cheaper (fewer reasoning tokens re-spent) and more coherent across a long session.