What's new in Chat GPT-4? June 2023
According to the OpenAI Help Center, the latest update for
ChatGPT was on June 22, 2023 and it included two new features for the mobile
app:
- Browsing: Plus users can now use Browsing to get
comprehensive answers and current insights on events and information that
extend beyond the model's original training data. To try it out, enable
Browsing in the “new features” section of your app settings. Then select GPT-4
in the model switcher and choose “Browse with Bing” in the drop-down.
- Search History Improvements: Tapping on a search result takes you directly to the respective point in the conversation.
More details:
You can also access the new GPT-4 models using the Chat Completion API, which is a new dedicated API for interacting with ChatGPT and GPT-4. GPT-4 is a new series of models that finished training in early 2023 and have improved performance and safety.
GPT-4 is the latest and most advanced system from OpenAI, which produces safer and more useful responses than ChatGPT. It is a large multimodal model that can accept text and image inputs and generate text outputs. Some of the details and capabilities of GPT-4 are:
- It can solve math problems, answer questions, make
inferences or tell stories, summarize large chunks of content, analyze photos
that users upload, make coding easier, provide more precise responses, and streamline
work across various industries.
- It can pass a simulated bar exam with a score around the
top 10% of test takers, and score in the 99th percentile on the Biology
Olympiad.
- It can generate, edit, and iterate with users on creative
and technical writing tasks, such as composing songs, writing screenplays, or
learning a user’s writing style.
- It has a maximum token count of 32,768, which translates
to around 64,000 words or 50 pages of text. This means it can keep up to 50
pages of context in mind when generating text.
- It is 82% less likely to respond to requests for
disallowed content and 40% more likely to produce factual responses than
GPT-3.5 on internal evaluations.
You can learn more about GPT-4 by reading the paper here:
https://openai.com/research/gpt-4 or by trying it on ChatGPT Plus here:
https://chat.openai.com/plus
However you have to pay for the tokens you use when you interact with the model. A token is a piece of word used for natural language processing, and 1,000 tokens is about 750 words. The price depends on the model and the context length you choose. For example:
- For models with 8k context length (e.g. gpt-4 and
gpt-4-0314), the price is: $0.03/1k prompt tokens and $0.06/1k sampled tokens.
- For models with 32k context lengths (e.g. gpt-4-32k and
gpt-4-32k-0314), the price is: $0.06/1k prompt tokens and $0.12/1k sampled
tokens.
You can find more details about the pricing here:
https://openai.com/pricing/
