OpenAI launches an upgraded large-scale language model called GPT-4 Turbo

Following the mid-March release of the groundbreaking GPT-4 language model by OpenAI, which features hybrid input capabilities and enhanced image recognition, the company recently unveiled an advanced iteration named GPT-4 Turbo at its inaugural developers’ conference.

Surpassing the previous GPT-4’s capacity to parse text content up to 50 pages, GPT-4 Turbo boasts the ability to comprehend as much as 300 pages of content, indicating its proficiency in deciphering longer and more intricate textual data. Moreover, GPT-4 Turbo’s training data is current up to April of this year, while the original GPT-4’s data was up to date only until September 2021. Nonetheless, it retains the capability to update real-time data via web browsing.

GPT-4 Turbo also maintains the ability to incorporate image content into the input interface, thereby generating image and text-based instructions and even handling text-to-speech conversions or direct uploads of document files for content analysis.

Additionally, it substantially reduces the cost of service input and output for developers, slashing expenses by up to threefold compared to earlier versions, thus enticing a broader developer base to utilize the new GPT-4 Turbo services.

OpenAI divulged that it currently has over 100 million active users weekly worldwide and is utilized by more than 92% of Fortune 500 companies. Furthermore, OpenAI has pledged to provide legal defense and cover the accompanying litigation costs should any copyright issues arise with the global use of their services.

GPT-4 Turbo is initially being offered in a preview mode, with no formal launch date for the official version disclosed as of yet. Regarding usage fees, specific costs for GPT-4 Turbo remain unpublished, while GPT-4 is currently available at a monthly subscription rate of $20.