Giving too long prompts won't yield good results from ChatGPT, says OpenAI's new guide
Giving too long prompts won't yield good results from ChatGPT, says OpenAI's new guide OpenAI 's new guide disproves the common belief that the longer and more detailed instructions (prompts) given to artificial intelligence (AI) models, the better the results. With the release of the latest version of Chat GPT , GPT-5.6 (GPT-5.6), the company has released a new guide on best prompting practices . According to this, OpenAI itself has confirmed the fact that only clear and simple prompts will give the best results. According to OpenAI's internal tests, removing unnecessary instructions and using simple prompts improves the evaluation scores of AI models by 10 to 15 percent. In addition, the company claims that total token consumption will decrease by 41 to 66 percent and user costs will decrease by 33 to 67 percent. Research has also shown that limiting Large Language Models (LLMs) to short answers produces more accurate results. Which model for which job? OpenAI has...