The advent of GPT-4.5: Available in 2024??

On the occasion of the first anniversary of GPT-4, clues on social networks and accidental leaks have suggested the imminent arrival of this new version! Version 4.5.

This article explores the contours of this evolution, the expectations it generates and the implications for the future of AI.

The online community was stirred up by several Twitter accounts and a short-lived post on the OpenAI website, announcing the release of GPT-4.5. This version, nicknamed “turbo”, promised improvements including a processing capacity extended to 256,000 tokens and a reduced cost compared to its predecessors. However, an unspecified hiccup appeared to delay this long-awaited release, leaving the community in suspense and speculation.

GPT-4.5 is positioned as a response to growing expectations in terms of artificial intelligence. With expanded context capability and significant cost reduction, this release promises to further democratize access to cutting-edge technologies. The only downside noted concerns the end date of knowledge of the model, set for June 2024, raising questions about its long-term relevance.

The Integration of GPT in the robotics world

The announcement of an impressive demo by OpenAI, in partnership with Figure, revealed the integration of ChatGPT into a robot.

This breakthrough illustrates the convergence between text-based artificial intelligence and robotics, paving the way for revolutionary practical applications. The robot, capable of understanding its environment and acting accordingly, marks a giant step towards machine autonomy.

The New Horizons of Apple and Google

Apple has quietly revealed the study of three new AI models, MM1, with high performance and specialization in chat, suggesting future integration into its devices. At the same time, Google DeepMind presented Agent Sima, capable of learning from gameplay videos, once again demonstrating the extent of the possibilities offered by AI in learning and interaction.

The Open Devin initiative, aimed at creating a 100% open source AI project, highlights the importance of collaboration and knowledge sharing in the field of AI. This approach could accelerate the development of technologies that benefit everyone, in contrast to the proprietary models that currently dominate the landscape.