Skip to main content

πŸ’¬ Responses

To an HTTP request, the API returns all results formatted in JSON.

Response format​

The response is made of three parts:

  • A request_id, uniquely identifying the request.
  • An outputs list, containing the model answer to your request, and useful metadata.
  • An execution_metadata dictionary, with the detailed cost of the request.

Let's have a detailed look at the request example generated on the Requests page.

Example response (JSON)
"request_id": "734e8d5a-a186-4d14-b2e4-be26f7fee6dc",
"outputs": [
"input_text": "Il Γ©tait une fois",
"completions": [
"output_text": ", un pays oΓΉ il faisait toujours beau.\nLes gens y vivaient heureux et en sΓ©curitΓ©. Ils avaient de l'argent",
"score": {
"logprob": -42.8507080078125,
"normalized_logprob": -1.7140283203125,
"token_logprobs": null
"execution_metadata": {
"cost": {
"tokens_used": 29,
"tokens_input": 4,
"tokens_generated": 25,
"cost_type": "lyra-fr@default",
"batch_size": 1
"finish_reason": "length"
"costs": {
"lyra-fr@default": {
"total_tokens_used": 29,
"total_tokens_input": 4,
"total_tokens_generated": 25,
"batch_size": 1

Common structures​

Execution metadata​

The execution_metadata dictionary collects information relevant to the cost and execution of the request. It is available at the top level, as well as for each individual element of a batch.

  • It contains a cost entry, which is a dictionary containing the detailed total cost of the request:

    • tokens_used, the number of tokens used (sum of the next two fields).
    • tokens_input, the number of tokens sent in input to the model.
    • tokens_generated, the number of tokens generated by the model.
    • cost_type in the form model_name@skill, indicating the nature of the tokens used (if no skills are used, it will be replaced by default).
    • batch_size the number of requests made in a single batch.
  • And a finish_reason entry, explaining why the model stopped processing further tokens (length if stopped by n_tokens or by reaching the end of the text to process, or stop_word if reached one of the stop_words).


The score dictionary provides information regarding the log-probabilities of the tokens processed:

  • logprob is the overall log-probability of the entire text processed.
  • normalized_logprob is the same as above, but normalized for text length (number of tokens).
  • token_logprobs is a dictionary including the specific log-probability of each token.

Response to batched requests​

The outputs list will be structured according to how you have batched your request:

  • It will contain one separate list for each set of parameters you have submitted.
  • Each list will contain one entry per entry in your batch.