update

Author: mrbullwinkle

articles/ai-services/openai/includes/api-versions/latest-inference-preview.md

@@ -42,9 +42,9 @@ Creates a completion for the provided prompt, parameters and chosen model.
 | best_of | integer | Generates `best_of` completions server-side and returns the "best" (the one with the highest log probability per token). Results cannot be streamed.<br><br>When used with `n`, `best_of` controls the number of candidate completions and `n` specifies how many to return `best_of` must be greater than `n`.<br><br>**Note:** Because this parameter generates many completions, it can quickly consume your token quota. Use carefully and ensure that you have reasonable settings for `max_tokens` and `stop`.<br> | No | 1 |
 | echo | boolean | Echo back the prompt in addition to the completion<br> | No | False |
 | frequency_penalty | number | Number between -2.0 and 2.0. Positive values penalize new tokens based on their existing frequency in the text so far, decreasing the model's likelihood to repeat the same line verbatim.<br> | No | 0 |
-| logit_bias | object | Modify the likelihood of specified tokens appearing in the completion.<br><br>Accepts a JSON object that maps tokens (specified by their token ID in the GPT tokenizer) to an associated bias value from -100 to 100. You can use this [tokenizer tool](https://platform.openai.com/tokenizer?view=bpe) to convert text to token IDs. Mathematically, the bias is added to the logits generated by the model prior to sampling. The exact effect will vary per model, but values between -1 and 1 should decrease or increase likelihood of selection; values like -100 or 100 should result in a ban or exclusive selection of the relevant token.<br><br>As an example, you can pass `{"50256": -100}` to prevent the <&#124;endoftext&#124;> token from being generated.<br> | No | None |
+| logit_bias | object | Modify the likelihood of specified tokens appearing in the completion.<br><br>Accepts a JSON object that maps tokens (specified by their token ID in the GPT tokenizer) to an associated bias value from -100 to 100. Mathematically, the bias is added to the logits generated by the model prior to sampling. The exact effect will vary per model, but values between -1 and 1 should decrease or increase likelihood of selection; values like -100 or 100 should result in a ban or exclusive selection of the relevant token.<br><br>As an example, you can pass `{"50256": -100}` to prevent the <&#124;endoftext&#124;> token from being generated.<br> | No | None |
 | logprobs | integer | Include the log probabilities on the `logprobs` most likely output tokens, as well the chosen tokens. For example, if `logprobs` is 5, the API will return a list of the 5 most likely tokens. The API will always return the `logprob` of the sampled token, so there may be up to `logprobs+1` elements in the response.<br><br>The maximum value for `logprobs` is 5.<br> | No | None |
-| max_tokens | integer | The maximum number of [tokens](https://platform.openai.com/tokenizer) that can be generated in the completion.<br><br>The token count of your prompt plus `max_tokens` cannot exceed the model's context length. [Example Python code](https://cookbook.openai.com/examples/how_to_count_tokens_with_tiktoken) for counting tokens.<br> | No | 16 |
+| max_tokens | integer | The maximum number of tokens that can be generated in the completion.<br><br>The token count of your prompt plus `max_tokens` cannot exceed the model's context length. [Example Python code](https://cookbook.openai.com/examples/how_to_count_tokens_with_tiktoken) for counting tokens.<br> | No | 16 |
 | n | integer | How many completions to generate for each prompt.<br><br>**Note:** Because this parameter generates many completions, it can quickly consume your token quota. Use carefully and ensure that you have reasonable settings for `max_tokens` and `stop`.<br> | No | 1 |
 | presence_penalty | number | Number between -2.0 and 2.0. Positive values penalize new tokens based on whether they appear in the text so far, increasing the model's likelihood to talk about new topics.<br> | No | 0 |
 | prompt | string or array | The prompt(s) to generate completions for, encoded as a string, array of strings, array of tokens, or array of token arrays.<br><br>Note that <&#124;endoftext&#124;> is the document separator that the model sees during training, so if a prompt is not specified the model will generate as if from the beginning of a new document.<br> | Yes |  |
@@ -289,7 +289,7 @@ Creates a completion for the chat message
 | logit_bias | object | Modify the likelihood of specified tokens appearing in the completion.<br><br>Accepts a JSON object that maps tokens (specified by their token ID in the tokenizer) to an associated bias value from -100 to 100. Mathematically, the bias is added to the logits generated by the model prior to sampling. The exact effect will vary per model, but values between -1 and 1 should decrease or increase likelihood of selection; values like -100 or 100 should result in a ban or exclusive selection of the relevant token.<br> | No | None |
 | logprobs | boolean | Whether to return log probabilities of the output tokens or not. If true, returns the log probabilities of each output token returned in the `content` of `message`. | No | False |
 | max_completion_tokens | integer | An upper bound for the number of tokens that can be generated for a completion, including visible output tokens and reasoning tokens. This is only supported in o1 series models. Will expand the support to other models in future API release. | No |  |
-| max_tokens | integer | The maximum number of [tokens](/tokenizer) that can be generated in the chat completion.<br><br>The total length of input tokens and generated tokens is limited by the model's context length. [Example Python code](https://cookbook.openai.com/examples/how_to_count_tokens_with_tiktoken) for counting tokens.<br> | No |  |
+| max_tokens | integer | The maximum number of tokens that can be generated in the chat completion.<br><br>The total length of input tokens and generated tokens is limited by the model's context length. [Example Python code](https://cookbook.openai.com/examples/how_to_count_tokens_with_tiktoken) for counting tokens.<br> | No |  |
 | messages | array | A list of messages comprising the conversation so far. [Example Python code](https://github.com/openai/openai-cookbook/blob/main/examples/How_to_format_inputs_to_ChatGPT_models.ipynb). | Yes |  |
 | metadata | object | Developer-defined tags and values used for filtering completions in the stored completions dashboard. | No |  |
 | modalities | [ChatCompletionModalities](#chatcompletionmodalities) | Output types that you would like the model to generate for this request.<br>Most models are capable of generating text, which is the default:<br><br>`["text"]`<br><br>The `gpt-4o-audio-preview` model can also be used to generate audio. To<br>request that this model generate both text and audio responses, you can<br>use:<br><br>`["text", "audio"]`<br> | No |  |
@@ -4887,7 +4887,7 @@ Information about the content filtering category (hate, sexual, violence, self_h
 | frequency_penalty | number | Number between -2.0 and 2.0. Positive values penalize new tokens based on their existing frequency in the text so far, decreasing the model's likelihood to repeat the same line verbatim.<br> | No | 0 |
 | logit_bias | object | Modify the likelihood of specified tokens appearing in the completion.<br><br>Accepts a JSON object that maps tokens (specified by their token ID in the GPT tokenizer) to an associated bias value from -100 to 100. You can use this [tokenizer tool](https://platform.openai.com/tokenizer?view=bpe) to convert text to token IDs. Mathematically, the bias is added to the logits generated by the model prior to sampling. The exact effect will vary per model, but values between -1 and 1 should decrease or increase likelihood of selection; values like -100 or 100 should result in a ban or exclusive selection of the relevant token.<br><br>As an example, you can pass `{"50256": -100}` to prevent the <&#124;endoftext&#124;> token from being generated.<br> | No | None |
 | logprobs | integer | Include the log probabilities on the `logprobs` most likely output tokens, as well the chosen tokens. For example, if `logprobs` is 5, the API will return a list of the 5 most likely tokens. The API will always return the `logprob` of the sampled token, so there may be up to `logprobs+1` elements in the response.<br><br>The maximum value for `logprobs` is 5.<br> | No | None |
-| max_tokens | integer | The maximum number of [tokens](https://platform.openai.com/tokenizer) that can be generated in the completion.<br><br>The token count of your prompt plus `max_tokens` cannot exceed the model's context length. [Example Python code](https://cookbook.openai.com/examples/how_to_count_tokens_with_tiktoken) for counting tokens.<br> | No | 16 |
+| max_tokens | integer | The maximum number of tokens that can be generated in the completion.<br><br>The token count of your prompt plus `max_tokens` cannot exceed the model's context length. [Example Python code](https://cookbook.openai.com/examples/how_to_count_tokens_with_tiktoken) for counting tokens.<br> | No | 16 |
 | n | integer | How many completions to generate for each prompt.<br><br>**Note:** Because this parameter generates many completions, it can quickly consume your token quota. Use carefully and ensure that you have reasonable settings for `max_tokens` and `stop`.<br> | No | 1 |
 | presence_penalty | number | Number between -2.0 and 2.0. Positive values penalize new tokens based on whether they appear in the text so far, increasing the model's likelihood to talk about new topics.<br> | No | 0 |
 | prompt | string or array | The prompt(s) to generate completions for, encoded as a string, array of strings, array of tokens, or array of token arrays.<br><br>Note that <&#124;endoftext&#124;> is the document separator that the model sees during training, so if a prompt is not specified the model will generate as if from the beginning of a new document.<br> | Yes |  |
@@ -4929,7 +4929,7 @@ Represents a completion response from the API. Note: both the streamed and non-s
 | logit_bias | object | Modify the likelihood of specified tokens appearing in the completion.<br><br>Accepts a JSON object that maps tokens (specified by their token ID in the tokenizer) to an associated bias value from -100 to 100. Mathematically, the bias is added to the logits generated by the model prior to sampling. The exact effect will vary per model, but values between -1 and 1 should decrease or increase likelihood of selection; values like -100 or 100 should result in a ban or exclusive selection of the relevant token.<br> | No | None |
 | logprobs | boolean | Whether to return log probabilities of the output tokens or not. If true, returns the log probabilities of each output token returned in the `content` of `message`. | No | False |
 | max_completion_tokens | integer | An upper bound for the number of tokens that can be generated for a completion, including visible output tokens and reasoning tokens. This is only supported in o1 series models. Will expand the support to other models in future API release. | No |  |
-| max_tokens | integer | The maximum number of [tokens](/tokenizer) that can be generated in the chat completion.<br><br>The total length of input tokens and generated tokens is limited by the model's context length. [Example Python code](https://cookbook.openai.com/examples/how_to_count_tokens_with_tiktoken) for counting tokens.<br> | No |  |
+| max_tokens | integer | The maximum number of tokens that can be generated in the chat completion.<br><br>The total length of input tokens and generated tokens is limited by the model's context length. [Example Python code](https://cookbook.openai.com/examples/how_to_count_tokens_with_tiktoken) for counting tokens.<br> | No |  |
 | messages | array | A list of messages comprising the conversation so far. [Example Python code](https://github.com/openai/openai-cookbook/blob/main/examples/How_to_format_inputs_to_ChatGPT_models.ipynb). | Yes |  |
 | metadata | object | Developer-defined tags and values used for filtering completions in the stored completions dashboard. | No |  |
 | modalities | [ChatCompletionModalities](#chatcompletionmodalities) | Output types that you would like the model to generate for this request.<br>Most models are capable of generating text, which is the default:<br><br>`["text"]`<br><br>The `gpt-4o-audio-preview` model can also be used to generate audio. To<br>request that this model generate both text and audio responses, you can<br>use:<br><br>`["text", "audio"]`<br> | No |  |