OpenAI launches GPT-4 Turbo model with latest knowledge cutoff

OpenAI announced a new model called GPT-4 Turbo at its first developer conference. It is superior in every way to the GPT-4 model and brings many new changes that developers and regular users have long requested. Additionally, the new model will be updated until April 2023 and will be available at an even lower price. If you want to learn more about OpenAI’s GPT-4 Turbo model, keep reading.

GPT-4 turbo model now available!

The GPT-4 Turbo model supports a maximum context window of 128K, which is even longer than Claude’s 100K context length. OpenAI’s GPT-4 model is now generally available for some users with up to 8K and 32K tokens. According to OpenAI, the new model can ingest more than 300 pages of books at a time, which is impressive.

Don’t forget, OpenAI has finally updated the knowledge cutoff for the GPT-4 Turbo model to April 2023 . On the user side, the ChatGPT experience has been improved and users can start using the GPT-4 Turbo model today. What’s amazing is that you don’t have to choose a specific mode to accomplish your task. ChatGPT now lets you smartly choose what to use when you need it. Browse the web, use plugins, analyze code, and more all in one mode.

A lot of new things have been announced for developers. First, the company launched a new text-to-speech (TTS) model . Produce incredibly natural voices with 6 different presets. Additionally, OpenAI is releasing the next version of its open source speech recognition model, Whisper V3 , which will soon be available via API.

Interestingly , APIs for Dall -E 3, GPT-4 Turbo with Vision, and new TTS models were released today. Coke today launches a Diwali campaign where customers can generate Diwali cards using the Dall -E 3 API. Next, there is a JSON mode that allows the model to respond with valid JSON output.

Additionally, the new model also improves function calls. OpenAI gives developers more control over their models . You can now set seed parameters to get consistent and reproducible output.

With the launch of tweak support, developers can now submit GPT-4 tweaks under the Experimental Access program. GPT-4 has been upgraded to a higher rate limit (twice the token per minute limit). Finally, when it comes to price, the GPT-4 Turbo model is significantly cheaper than the GPT-4. The cost is 1 cent for 1,000 input tokens and 3 cents for 1,000 output tokens. Effectively, GPT-4 Turbo is 2.75 times cheaper than GPT-4 .

So, what do you think about the new GPT-4 Turbo model? Let us know in the comments section below.