The tests have shown that 33%of the answers to the questions about O3, which are twice as higher than O1 (16%) and O1 (16%) and O3-Mini models (14.8%) (14.8%). O4-Mini model turned out to be less accurate, 48% of cases “hallucination”.
The Transoud Independent Laboratory claimed that O3 sometimes claimed that it launched a code in the 2021 MacBook Pro, which is technically impossible, technically impossible. Openai does not yet understand why new models are more often wrong, which may be related to the teaching method – strengthening learning (strengthening learning). This complicates the use of models in areas where accuracy is critical, for example in case -law.
One of the solutions is the integration of web mail, as in the GPT-4O, which reaches 90% of the accuracy in the Simpleqa test. Openai continues to investigate to reduce error levels.
Source: Ferra
I am a professional journalist and content creator with extensive experience writing for news websites. I currently work as an author at Gadget Onus, where I specialize in covering hot news topics. My written pieces have been published on some of the biggest media outlets around the world, including The Guardian and BBC News.
