During its first-ever developer conference on Monday, OpenAI previewed GPT-4 Turbo, a brand new version of the large language model that powers its flagship product, ChatGPT. The newest model is capable of accepting much longer inputs than previous versions — up to 300 pages of text, compared to the current limit of 50. This means that theoretically, prompts can be a lot longer and more complex, and responses might be more meaningful.
OpenAI has also updated the data that GPT-4 Turbo is trained on. The company claims that the newest model now has knowledge about the world until April 2023. The previous version was only caught up until September 2021, although recent updates to the non-Turbo GPT-4 did include the ability to browse the internet to get the latest information.
GPT-4 Turbo will also accept images as prompts directly in the chat box, wherein it can generate captions or provide a description of what the image depicts. It will also handle text-to-speech requests. And users will now be able to upload documents directly and ask the service to analyze them — a capability that other chatbots like Anthropic’s Claude have included for months.
For developers, using the newest model will effectively be three times cheaper. OpenAI said that it was slashing costs for input and output tokens — a unit used by large language models to understand instructions and respond with answers.
In addition to announcing its newest large language model, OpenAI revealed that ChatGPT now has more than 100 million weekly active users around the world and is used by more than 92 percent of Fortune 500 companies. The company also said that it would defend customers, including enterprises, not only against legal claims around copyright infringement that might arise as a result of using its products, but it would also pay for costs incurred as a result.
OpenAI also revealed single-application “mini-ChatGPTs” today, small tools that are focused on a single task that can be built without even knowing how to code. GPTs created by the community can be immediately shared, and OpenAI will open a “store” where verified builders can make their creation available to anyone.
The company didn’t announce when GPT-4 Turbo would come out of preview and be available more generally. Accessing GPT-4 currently costs $20 a month.
This article originally appeared on Engadget at https://www.engadget.com/gpt-4-turbo-is-openais-most-powerful-large-language-model-yet-211956553.html?src=rss