GPT-6 Astra Migration Guide

​

What’s New

• Mid-conversation effort with configuration_update: change reasoning effort from a point in the conversation onward without invalidating the prompt cache

​

Mid-Conversation Effort with configuration_update

​

Responses API

{
  "model": "openai/gpt-6-astra",
  "reasoning": { "effort": "medium" },
  "input": [
    { "role": "user", "content": "Summarize the incident report." },
    {
      "type": "message",
      "role": "assistant",
      "content": [{ "type": "output_text", "text": "The outage began at 09:14 UTC when ..." }]
    },
    { "type": "configuration_update", "reasoning": { "effort": "high" } },
    { "role": "user", "content": "Now find the root cause and propose a fix." }
  ]
}

​

Chat Completions API

{
  "model": "openai/gpt-6-astra",
  "reasoning": { "effort": "medium" },
  "messages": [
    { "role": "user", "content": "Summarize the incident report." },
    { "role": "assistant", "content": "The outage began at 09:14 UTC when ..." },
    {
      "role": "system",
      "content": "",
      "configuration_update": { "reasoning": { "effort": "high" } }
    },
    { "role": "user", "content": "Now find the root cause and propose a fix." }
  ]
}

​

Anthropic Messages API

{
  "model": "openai/gpt-6-astra",
  "max_tokens": 4096,
  "output_config": { "effort": "medium" },
  "messages": [
    { "role": "user", "content": "Summarize the incident report." },
    { "role": "assistant", "content": "The outage began at 09:14 UTC when ..." },
    { "role": "system", "content": [], "output_config": { "effort": "high" } },
    { "role": "user", "content": "Now find the root cause and propose a fix." }
  ]
}

​

Effort values

​

Placement rules

• Put the update directly before the user turn it should apply to. An update ahead of the first user turn is allowed.
• Do not place two updates next to each other. Adjacent updates are rejected.
• A request must not end with an update. OpenRouter rejects this with a 400 because the update has no following input item to apply to.
• Updates cannot be combined with truncation: "auto" in the Responses API. Automatic truncation could drop the update and silently revert the effort.
• Keep the update at the same position in later requests. Moving or removing it changes the prefix and loses the cache benefit.
• Mid-conversation effort updates are not accepted on the Batch API and are rejected per line.

​

Prompt caching

​

Migration Checklist

1. Swap the model slug to openai/gpt-6-astra.
2. Keep your existing request-level reasoning.effort as the baseline. Do not change it per turn.
3. Where one turn needs more or less reasoning, insert a configuration_update (Responses), a content-less system message carrying configuration_update (Chat Completions), or a content-less system message carrying output_config.effort (Messages) directly before that user turn.
4. Keep every update at its original position when you replay history, and never place two updates back to back or end a request with one.
5. Remove truncation: "auto" from Responses requests that carry updates.
6. Watch cached_tokens in usage to confirm the prompt prefix is still reused.

​

Breaking Changes

• Requests without an update behave the same as before
• The request-level reasoning.effort keeps its meaning as the conversation baseline
• Other models and other OpenAI Responses transports are unaffected. A request that carries an update is routed only to endpoints that accept it

​

Resources

• Using GPT-6 Astra, OpenAI’s model guidance with migration and prompting best practices for GPT-6 Astra
• Change reasoning mid-conversation in OpenAI’s reasoning guide
• Changing Effort Mid-Conversation in the OpenRouter reasoning guide
• Reasoning Tokens
• Prompt Caching
• OpenRouter Responses API