OpenAI launched GPT-4 in March this year, but only invited some developers who had submitted applications to participate in the test; In a press release today, OpenAI announced that GPT-4 is open to all developers who pay for API access.
GPT-4 is another major breakthrough after GPT-3, with more than 100 billion parameters, which is 10 times the number of GPT-3. GPT-4 can generate natural language output of various types and styles, such as articles, dialogues, abstracts, poems, lyrics, etc., based on a given text or speech input.
OpenAI also said that more apis including GPT-3.5 Turbo, Whisper and its DALL · E image generation will be available starting today.
The blog post says OpenAI is working on “implementing fine-tuning of GPT-4 and GPT-3.5 Turbo” and plans to make these services available to developers later in 2023.
According to the OpenAI blog post, the opening of GPT-4 is an important step in OpenAI’s vision of providing universal Artificial Intelligence (AGI) capabilities and opportunities for all.
According to OpenAI, GPT-4 can help developers and enterprises in education, entertainment, healthcare, business and other industries to create more innovative and valuable applications and services.
At the same time, OpenAI also emphasizes the safety and ethical monitoring and management of GPT-4 to prevent its abuse or misuse.
Foreign science and technology media NeoWin believes that the opening of GPT-4 is a milestone event in the field of artificial intelligence, but also a huge boost to natural language processing technology. GPT-4 not only demonstrates its powerful generative power and flexibility, but also offers more possibilities and potential for communication and cooperation between humans and machines.