#look at this. i put these responses into a word counter, and it comes out to only 4,818 graphemes an

1 messages · Page 1 of 1 (latest)

tired cargo
inner falconBOT
#

Hi @tired cargo. I'm kapa.ai, your friendly support bot. I will try to answer your question by searching through the knowledge base. :wave:

#

The number of graphemes (characters or visible symbols) in a text is not the same as the number of tokens used by OpenAI models. Tokens are the units that language models process, and they do not map 1:1 to graphemes or words. In English, a token can be as short as one character or as long as one word, but on average, 1 token is about 4 characters or 0.75 words. This means that 4,818 graphemes could correspond to a different number of tokens, depending on the specific text and language.

For example, the string "ChatGPT is great!" is encoded into six tokens: ["Chat", "G", "PT", " is", " great", "!"], even though it has 17 characters (graphemes) Managing tokens.

If you want to know exactly how many tokens your text will use, you should use a tokenizer tool like tiktoken or the OpenAI Tokenizer web tool. This will give you an accurate token count, which is what matters for API usage and billing—not the grapheme or word count How to count tokens with Tiktoken.

In summary:

  • Grapheme count and token count are not equivalent.
  • Token count determines API usage and cost.
  • Use a tokenizer to get the exact token count for your text.

Note: you can ask me a follow up question by @ mentioning me again :speech_balloon:

#

@tired cargo Kapa.ai is still learning and improving, please let me know how I did by reacting below.

tired cargo
inner falconBOT
#

Let me check the knowledge base for you again. One moment! :books:

#

I'm sorry, but the knowledge sources provided do not contain information about changes to the maximum response length (in graphemes, tokens, or words) for ChatGPT responses between February 2025 and now. There is no explicit mention of a reduction in response length or any policy change regarding output size in the official release notes or documentation.

If you are experiencing noticeably shorter responses, it may be due to ongoing model updates, changes in model behavior, or temporary system adjustments, but this is not documented in the sources available. For the most accurate and up-to-date information, you may want to reach out to OpenAI support directly or monitor the official release notes for any future updates.

Note: you can ask me a follow up question by @ mentioning me again :speech_balloon:

#

@tired cargo Kapa.ai is still learning and improving, please let me know how I did by reacting below.