[Additive] OpenAI: 4 new response fields

OpenAI2 breaking4 additive

• Breakingdirection flip: their tools call this safeprompt_cache_options is no longer acceptedUsually ignored rather than rejected, so the call succeeds and does something subtly different. 2 places: CreateModelResponseProperties.prompt_cache_options, BetaCreateModelResponseProperties.prompt_cache_options## Code that breaks

await client.post("/…", {
  // prompt_cache_options is now ignored, not rejected
  prompt_cache_options: value,
});

• Breakingdirection flip: their tools call this safevalue can now return "max_output_tokens"Producer-side tools call this additive. It breaks exhaustive switches, strict parsers and generated client enums. SessionTurnErrorCodeResource · reachable from 6 operations## Code that breaks

switch (response) {
  // existing branches …
  // no branch for "max_output_tokens"
  default:
    throw new Error("unexpected value");
}

4 new response fields #

• ResponseUsage.input_tokens_details.cache_write_tokens_details
• BetaResponseUsage.input_tokens_details.cache_write_tokens_details
• BetaResponse.usage.input_tokens_details.cache_write_tokens_details
• Response.usage.input_tokens_details.cache_write_tokens_details