The article claims that ChatGPT 3.5 users can access the 4.0 Plus features for free via a link. It provides the link and shows that after logging in through the link, users can access Plus features like GPT-4 and DALL-E. The article speculates that this could be a deliberate leak by OpenAI to manage the pressure from increased usage. It suggests readers take advantage of the opportunity to try out Plus features for free, as it is unclear when the vulnerability will be fixed. The article also mentions that OpenAI's CEO recently announced a pause on new user registrations.

iFlytek announced the release of the Spark Cognition Model V3.0, surpassing ChatGPT. Spark 3.0 marks the evolution of general models, aiming to match ChatGPT 3.5, with superior performance in Chinese and comparable performance in English. The seven capabilities of text generation, language understanding, knowledge Q&A, logical reasoning, mathematical skills, coding abilities, and multimodal abilities have all seen continuous improvements.

Anthropic's Claude AI model excelled in performance evaluations, demonstrating strong results even in its worst version. The Claude model shows advantages in handling large-scale contextual inputs and long prompts, sparking discussions around the applications of AI chatbots. The importance of open-source models has been highlighted in this competition, with Meta's Llama2LongAI also making significant strides. The success of the Claude model could influence the decisions of businesses and users in selecting AI models.