OpenAI takes the heat out of AI with the modern GPT-4 and cheap GPT-3.5 Turbo.

OpenAI has released new models, including an updated GPT-4 Turbo preview model, and reduced access to the GPT-3.5 Turbo application programming interface (API). The company also introduced new ways for developers to manage API keys and support API usage.

In a blog post, OpenAI said the updated GPT-4 Turbo “completes tasks like code generation more than the previous preview model and is intended to reduce the “laziness” of the model not completing a task.”

Introducing two more model upgrades:

• Improved GPT-4 Turbo preview, improves functions such as code generation and fixes bug in UTF-8 encoding • New GPT-3.5 Turbo next week, input cost reduced by 50% and output cost reduced by 25%https://t.co/mNGcmLLJA8

— Open AI Devs (@OpenAIDevs) January 25, 2024

OpenAI also said that it is introducing a new GPT-3.5 Turbo model, gpt-3.5-turbo-0125, and for the third time in the past year, it is reducing the price on GPT-3.5 Turbo to help its customers. The input prices of the new model will be reduced by 50% to $0.0005 per thousand tokens, and the output prices will be reduced by 25% to $0.0015 per thousand tokens.

In December 2023, there were complaints among ChatGPT users about the chatbot's often-decreasing functionality, attributed to a lack of updates to GPT-4. While users of GPT-4 using data prior to September 2021 may still experience similar laziness issues, GPT-4 Turbo was recently trained on April 2023 data.

Related: OpenAI CEO Sam Altman sees funding for global chip venture

OpenAI introduces smaller AI models known as embeddings. OpenAI defines an embedding as a sequence of numbers that represent concepts in content such as language or code.

Embedding is a type of AI tool that helps computers understand and use written text more effectively. They do this by converting words and sentences into a format that computers can process. Think of embedding as a translator that converts human language into unique code that computers can understand and process.

Recovery-augmented generation is a type of AI that provides accurate and appropriate responses rather than generating responses from scratch. It's the kind with AI that quickly searches a reference book and tells you what to find instead of guessing the answer.

Two new models that use these embeddings are currently available: “text-embedding-3-small” and a more powerful version “text-embedding-3-large”. “Small” and “large” indicate the capacity of these models. The large model is like a general translator – it can understand and transform the text in a more sophisticated way than the “small”. These models are now available for applications seeking to extract and use data efficiently.

Simply put, these new tools are like smarter and more efficient computer translators, helping them better understand human language and quickly find the information they need from large databases. This results in more accurate and helpful responses when interacting with AI systems.

OpenAI's GPT-4 has competition from other artificial intelligence (AI) models such as Google's Gemini. Gemini beats GPT-4 by being able to do advanced math and special coding. However, some have argued that the results may have been different if the advanced Gemini model had been tested against the GPT-4 Turbo.

OpenAI also plans to introduce a way for GPT creators to monetize their personalized AI systems. US builders are paid by consumers with GPTs as a first step. However, the GPT storage is first rolled out to users with paid ChatGPT plans.

Magazine: Crypto+AI token picks, AGI will take ‘longer', Galaxy AI up to 100M phones: AI Eye