Rosalind Evans
In a groundbreaking announcement, OpenAI has supercharged its GPT-4 model, introducing GPT-4 Turbo, an API preview that boasts cost reductions and expanded functionalities for developers. The new model, enriched with data up to April 2023, is set to revolutionize the landscape of large language models with its affordability and updated knowledge bases.
OpenAI’s first developer conference was abuzz with the news that GPT-4 Turbo will be three times more affordable, with input costs slashed to $0.01 per 1,000 tokens and outputs priced at $0.03 per 1,000 tokens. This pricing strategy significantly undercuts the $0.03 input cost of the previous GPT-4 iteration. “GPT-4 Turbo is cheaper to run for developers,” the company highlighted, making it an attractive option for those looking to integrate advanced AI into their applications.
The improved GPT-4 Turbo also features a 128K context window, allowing it to process over 300 pages of text in a single prompt. This enhancement enables the model to deliver more nuanced and comprehensive responses. “GPT-4 Turbo will also ‘see’ more data,” OpenAI remarked, emphasizing the model’s sophisticated understanding and response capabilities.
Excitingly, GPT-4 Turbo retains the ability to accept image prompts and text-to-speech requests and integrates DALL-E 3, showcasing OpenAI’s commitment to multimodal AI solutions. Moreover, the model allows users to specify coding languages such as XML or JSON, catering to diverse development needs.
Improvements are not limited to GPT-4 Turbo. The GPT-3.5 Turbo model also receives upgrades in context window size and functionality, along with more competitive pricing. OpenAI’s initiative aligns with their earlier release of GPT-3.5 Turbo, which was touted as the optimal model for non-chat applications.
In an unprecedented move, OpenAI introduces Copyright Shield, providing copyright indemnity akin to that of industry giants Microsoft and Google. This program ensures OpenAI backs enterprise users in the face of copyright infringement claims. “We will now step in and defend our customers and pay the costs incurred if you face legal claims around copyright infringement,” the company assured.
As OpenAI continues to lead the charge in AI innovation, GPT-4 Turbo stands out as a testament to the company’s dedication to accessibility, efficiency, and legal security for its customers.
The evolution of OpenAI’s language models, notably GPT-4 Turbo, marks a new era in AI, where advanced capabilities become increasingly accessible to developers and businesses alike. With the integration of Copyright Shield, OpenAI not only propels the technological frontier forward but also ensures its users navigate this new terrain with confidence and support. The anticipation for the full release of GPT-4 Turbo is palpable, as it promises to usher in a wave of transformative AI applications.
