look news india
News India Live, Digital Desk: Openai Has launched the new version of its logic-focused AI model O3-PRO. It started on 10 June and is now available to the Chatgpt Pro and Team users, which is replacing the previous O1-PRO. According to the company, enterprise and education users will have access to next week.
The O3-PRO is based on the O3 model released earlier this year. It is designed to be slow, but more reliable, especially in subjects such as coding, mathematics and science. Openai says that this model performs better in prolonged thinking tasks, making it more useful in areas that require more accuracy at speed.
Designed for challenging questions
The biggest feature of O3-PRO is that it can deal more effectively than complex problems than previous models. Openai “recommends using for challenging questions where reliability matters more than speed.” This model can do everything that can do O3, including searching on the web, reading uploaded files, understanding images, running the pythan code and personalizing answers using memory.
It also means that it is slow. The model takes longer to respond to the O1-PRO, but its advantage is that it gives better answers. In internal reviews, Openai examiners preferred O3-PRO in every major field including science, programming, business questions and even writing aid.
Benchmarks put it ahead of rivals
OpenAI claims that the O3-PRO is now performing better than its main rivals. AIME 2024, on a Mathematics Benchmark, O3-PRO allegedly scored more than Google’s Gemini 2.5 Pro. It also defeated Anthropic’s Cloud 4 Opus in the GPQA Diamond Test, which examines the science understanding of PhD-level.
The company also tested the O3-PRO, which it calls the “4/4 credibility” method, in which AI has to give the correct answer to all four times when asked the same question. According to Openai, the O3-PRO was continuously successful in this test.
Openai O3-PRO Benchmark
Pricing and limitations
In API, the O3-PRO is priced at $ 20 per million input tokens and $ 80 per million output token. For reference, a million input tokens are around 7.5 million words. In Indian currency, it is for about ₹ 1,740 per million input tokens and ₹ 6,960 output.
It is said that this model is not correct. Openai says that temporary chats for O3-PRO are currently disabled due to technical disturbances. It cannot even create an image and users using Openai’s workspace tool canvas must switch to another model.
what comes next?
For users on ChatjPT Pro, the O3-PRO is now available directly from the model picker. If you are interested in writing material with long-for research, coding aid, or better accuracy, this model may be worth waiting. Just be ready for a little long response time.
OpenAI has said that the safety measures used in O3 also apply to O3-PRO.
Yunus’s ‘anti -India’ move in Bangladesh failed, army exposed