OpenAI has announced several enhancements to its leading large language models, GPT-4 and GPT-3.5, at the company’s inaugural developer conference. The upgraded GPT-4 Turbo has been trained on data through April 2023, expanding its knowledge base significantly from the previous September 2021 cut-off. This new version also increases the context window to 128,000 tokens, over 300 pages of text, enabling more nuanced responses.
GPT-4 Turbo also vastly increases the context window to 128,000 tokens, equivalent to over 300 pages of text in a single prompt. This allows the model to fully understand prompts and offer remarkably nuanced, relevant responses. Pricing has been slashed to just $0.01 per 1,000 tokens for input and $0.03 for output, making GPT-4 Turbo 3 times more affordable than previous iterations.
GPT-4 Turbo retains capabilities introduced earlier this year like image prompts, text-to-speech functionality, and integration with the DALL-E 3 image generation model. Users can leverage these features to get even more value from the upgraded model.
GPT-3.5 Turbo Receives Similar Enhancements Focused on Context and Cost
OpenAI’s GPT-3.5 Turbo, launched in March as the premiere model for non-conversational use cases, is also gaining expanded context windows up to 16,000 tokens. This allows for remarkably detailed responses for tasks like searching, summarizing, and translating. The pricing has been reduced to an extremely affordable $0.01 per 1,000 tokens for input and just $0.002 for output.
These upgrades make GPT-3.5 Turbo even better suited for developers looking to imbue applications with next-generation AI capabilities without breaking the bank. The model can now do even more within a single prompt.
New Copyright Protections Give Users Peace of Mind
Mirroring recent moves by competitors Microsoft and Google, OpenAI has launched the Copyright Shield program to protect customers from copyright lawsuits related to its AI services. OpenAI says it will fully cover any costs if users face legal action over alleged copyright infringement stemming from ChatGPT Enterprise or its developer platform.
This shields developers and companies from the risks of generative AI so they can focus on creating amazing new applications powered by models like GPT-4 Turbo and GPT-3.5 Turbo. OpenAI is at the forefront of providing robust protections along with its groundbreaking AI capabilities.