GPT-5.2 Still Can't Beat Gemini 3 in Simple Tests

GPT-5.2 was released in a hurry to compete with Apple's Gemini 3/DOK

JAKARTA - OpenAI launched the GPT-5.2 update on December 11, just a month after the release of the previous version. The release date that was less than satisfactory was considered rushed.

In the tests conducted by PC Mag, GPT-5.2 was called 'immature' because it was pulled out before its time. This hasty launch is also suspected to have been deliberately carried out by OpenAI to balance the dominance of the Gemini 3 model which is also newly released by Google.

This alleged 'fear of losing competitiveness' does not only arise because of its release date. When compared in standard testing, GPT-5.2 looks less superior in a number of things compared to Gemini 3.

Although OpenAI claims its latest Artificial Intelligence (AI) model is superior in making presentations and recognizing images, the results of GPT-5.2 work are still very similar to GPT-5.1. This makes the improvement in GPT-5.2 not so noticeable by users.

In testing the creation of spreadsheets containing video game data, GPT-5.2 is indeed a little more accurate than its predecessor. However, minor errors in the previous version can still be quickly corrected through normal conversations so that this update is not very noticeable.

On the other hand, Gemini 3 shows a much more noticeable improvement. Gemini 3 Flash is now much more accurate in recognizing the details of complex image components. In addition, compared to GPT-5.2, the work of its creative writing ability is much more unique.

The tight competition between OpenAI and Google is indeed beneficial for consumers because technological innovation becomes faster. However, OpenAI's very hasty strategy can lower the quality standards of the products they release to the public.