UNIT.City — місце, де люди працюють... КРАЩЕ! Обирай свій простір просто зараз 👉
Олександр КузьменкоAI Eng
6 December 2024, 13:54
2024-12-06
OpenAI announced a new subscription with "unlimited" access and a "biting" price
OpenAI’s new ChatGPT subscription plan provides unlimited access to all of its AI models, including the upgraded full version of o1 pro. However, the cost of ChatGPT Pro significantly hits the wallet and is 10 times higher than the price of ChatGPT Plus.
OpenAI’s new ChatGPT subscription plan provides unlimited access to all of its AI models, including the upgraded full version of o1 pro. However, the cost of ChatGPT Pro significantly hits the wallet and is 10 times higher than the price of ChatGPT Plus.
As reported in OpenAI, the new subscription level will cost users $200 per month.
«We think the audience for ChatGPT Pro will be advanced ChatGPT users—those who are already pushing models to their limits in tasks like math, programming, and writing,» said Jason Wei, a member of the OpenAI technical team.
Unlike most AI, o1 and other «thinking» models try to check their own work as it goes. This helps them avoid some of the mistakes that usually lie in wait for models, but they often take longer to find a solution. Model o1 «thinks» about the task, planning ahead and performing a series of actions that help it to suggest answers.
OpenAI released a preview of o1 in September, but the new version is generally more productive. Compared to the preview, users can expect «a faster, more powerful, and more accurate reasoning model that does even better with coding and math», — an OpenAI spokesperson said.
In addition, o1 can now «think» about loading images (this was not possible in the preview version), and has been trained to be «more concise in its thinking» to improve response time. According to OpenAI’s internal testing, o1 reduces the number of «basic errors» in «complex real-world questions» by 34% compared to the previous version.
However, surprisingly, the full version of o1 performs worse than the preview according to several popular benchmarks. One such test is the MLE-Bench, which measures how well AI «agents» do with machine learning engineering.
o1 does not require a maximum ChatGPT Pro subscription. All paid ChatGPT users can access it through the ChatGPT Model Picker.
What ChatGPT Pro does
ChatGPT Pro subscribers will get a supposedly better version of o1 than users who pay less. The o1 pro mode «uses more computing resources to get the best answers to the most complex questions», according to OpenAI.
ChatGPT Pro users can access this feature by selecting «o1 pro mode» in the model selection window and asking questions directly. Since responses will take longer to generate, ChatGPT will display a progress bar and send an in-app notification if the user switches to another conversation.
o1 pro mode can increase the model’s «thinking» time before it responds. The company previously said it would experiment with o1 models that «think» for hours, days, or even weeks to further enhance their «reasoning» ability.
«Compared to o1 and o1-preview, o1 pro mode performs better on complex machine learning tests in math, science, and coding. In particular, we saw a 75% reduction in errors in the simpler Coding Olympiad questions, which are more in line with everyday programming questions,» says an OpenAI representative.
ChatGPT Pro is OpenAI’s most expensive plan and is 10 times the cost of ChatGPT Plus. Experts believe that it will be difficult for the company to sell such a subscription, given that many people consider even ChatGPT Plus to be too expensive.
To make ChatGPT Pro more attractive, OpenAI also provides unlimited access to GPT-4o and advanced voice mode, a feature of ChatGPT similar to human conversation.
In September, it was reported that NASA research scientist Kyle Cabasares was testing a new artificial intelligence model from OpenAI, known internally as Strawberry, or generically o1. Using ChatGPT o1, he wrote the code for his PhD thesis in just 1 hour. Although, the researcher and a group of authors once spent about 10 months on this task.
Як нейромережі бачать вільну та незалежну Україну? Тест dev.ua
Нейронні мережі для генерації зображень бачать світ по-своєму, їхню логіку зрозуміти часом зовсім неможливо. Але таки хочеться. На честь Дня Незалежності України редакція dev.ua вирішила провести невеликий експеримент.
Ми задали чотирьом різним нейронним мережам п’ять однакових запитів: «прапор України», «День Незалежності України», «український Крим», «перемога України» та «українці». Отриманими результатами ми ділимося з вами нижче.
У TikTok тепер можна генерувати фон за допомогою нейромережі. Ми протестували її та ділимося результатами
У TikTok з’явилася нова функція «Розумний фон». З її допомогою як фон для тіктоків можна підставляти згенеровані нейромережею зображення. Редакція dev.ua протестувала цю технологію і ділиться своїми враженнями.