The latest advancements from the tech giants, GPT-5 and Llama-3 from OpenAI and Meta respectively will operate on the foundation of LLMs. They are scheduled for release before the end of summer 2024, but what can we expect from these new chatbots?

OpenAI’s GPT-5

Building on the strengths of its predecessor, GPT-4, the new chatbot will have a number of enhancements centred around user experiences and energy efficiency. 

The first of these upgrades will be multilingual proficiency: GPT-5 will help non-English speakers in translation tasks. Beyond simple translation, the upgrade will have improved abilities to interpret the nuances of language. 

GPT-5 will also have better reasoning abilities, making it able to handle more complex datasets and problems. It will be able to process up to 50,000 words at a time, twice as many as GPT-4. 

It will also have a higher level of emotional intelligence, with more empathetic interactions which may lead to better customer service value. 

OpenAI also claims GPT-5 will introduce more robust security protocols, with better defences against malicious exploitation and safeguarding user interactions. 

Finally, the new version will be able to integrate seamlessly with IoT devices, supporting the concept known as industry 5.0, where humans and machines operate interactively in the same workplace. 

Meta’s Llama-3

Meta’s response to the evolving landscape of AI comes in the form of Llama-3, with a number of notable attributes like an increased parameter count, scaling up to 400 billion, enhancing its predictive prowess.

It will also have proficiency in multiple languages and an expanded context window, allowing for quicker textual data processing. In short, this means the model will be able to handle larger chunks of text or data in a shorter period of time. 

Notably, the new version of Meta’s chatbot will be able to process text, images, and video. Meaning it will be able to take images as input and describe the content, but also make images from user prompts. Finally Llama-3 will encompass various application integrations, like Google Cloud. 

Recommended reading

• Meta To Start Labelling AI-generated Images
• OpenAI & Meta to Release AI Models that Can ‘Plan’ and ‘Reason’
• Meta Releases “Code Llama,” a Large Language Model for Coding

These advancements mark a transformative era in human-machine interaction, which continues to revolutionise industries and societies worldwide. 

From enhanced communication and healthcare delivery to increased efficiency in business operations, GPT-5 and Llama-3 continue towards a future where AI augments human capabilities.