ZDNET recently tested OpenAI’s latest chatbot model, GPT-5.2. The test aimed to evaluate the AI’s performance in various tasks, including text summarization, academic concept explanations, and image generation. Here are the results:
GPT-5.2 showed promise in some areas, such as text analysis and writing. However, it fell short in others, like coding and translation tests.
The model performed well in text-based tests, earning a score of 92 out of 100. In image-based tests, GPT-5.2 scored 17 out of 20 points.
One notable aspect of the model’s performance is its tendency towards brevity, which may be beneficial for users but also leads to frustration. The AI occasionally delays responses by several minutes, and some coding errors were reported in testing.
While GPT-5.2 represents progress over previous models, it feels more like an incremental update rather than a significant leap forward. As with any new technology, there is always room for improvement, and users can expect to see further refinements and updates from OpenAI in the future.
Source: https://www.zdnet.com/article/text-image-tests-openai-gpt-5-2-mixed-results