ChatGPT’s parent company OpenAI recently held its developer conference for the first time. During the event, a lot of announcements were made by OpenAI. It was during this event that OpenAI also shared its future plans for ChatGPT. Its large language models, both GPT-4 and GPT-3.5 will see major upgrades. Let’s find out more about this.
During the event, the CEO of OpenAI, Sam Altman announced the GPT-4 Turbo. They revealed that GPT-4 which had access to information only up to September 2021, will be succeeded by GPT-4 Turbo which will have information till April 2023. GPT-4 Turbo is currently available via API but OpenAI plans on releasing a product-ready Turbo version in the coming few weeks.
Also read: OpenAI’s Red Teaming Network deployed for the development of its models
Further, the GPT-4 Turbo will have access to a 128K context window. For reference, this is close to going through 300 pages in a single prompt. This will also ChatGPT to understand the question better and offer a more thoughtful answer to the prompts. GPT-4 was previously offered in two variants – 8K and 32K.
What can you do on GPT 4 Turbo? For starters, you can put in image prompts, even give text-to-speech requests, and even integrate DALL-E 3. You can even ask the chatbot to perform difficult tasks in a single prompt. Coding and much more can be done here as well.
Also read: OpenAI, Meta and Tesla desperately want this Nvidia GPU, but why?
And you might think that all of this might come for a lot of money. Well, GPT-4 Turbo will come at a cheaper price for developers. According to OpenAI, the new version of GPT4 is three times cheaper than the previous ones. Input for GPT-4 Turbo would cost $0.01 per 1,000 tokens (the basic unit of text or code for LLMs to read) compared to the $0.03 on GPT-4. So, each output will be $0.03 per 1,000 tokens.
