According to BGR , despite speculation about the appearance of GPT-5, OpenAI has just announced the new generation of reasoning model o3 which will be the 'heart' of ChatGPT in the future.
OpenAI's New Language Model Set to Launch Next Month
Closing out the '12 Days of ChatGPT' announcement series, OpenAI surprised everyone with o3, a model that promises to outperform its predecessors. A stripped-down version, o3-mini, will be released at the end of January, followed by the full o3.
OpenAI is about to release a new o3 language model
PHOTO: IXBT SCREENSHOT
However, o3 is not the long-awaited GPT-5. According to the Wall Street Journal , the GPT-5 project has been plagued by difficulties and delays. Known by the code name Orion, GPT-5 has been in development for 18 months and costs a whopping $500 million per run to train.
The main reason for the delay was the lack of high-quality data. Public internet data sources were exhausted, forcing OpenAI to look for alternatives such as synthetic data generation or hiring people to manually generate data. In addition, the training process for GPT-5 was technically challenging. Initial test runs showed that training GPT-5 would take much longer and cost more than expected.
The departure of many senior staff after internal turmoil at OpenAI in late 2023 is also said to be one of the reasons why the project was affected.
Meanwhile, o3 is expected to be an important step forward in the development of generative AI, with its superior reasoning capabilities. Will o3 be the key to helping OpenAI solve the data problem for GPT-5? The answer remains open.
Currently, OpenAI has not made any official announcement about the launch date of GPT-5.
Source: https://thanhnien.vn/openai-sap-ra-mat-mo-hinh-o3-gpt-5-van-bat-vo-am-tin-185241225090546797.htm
Comment (0)