En İyi Yapay Zekâ Modeli Araştırıldı: Şampiyon Ne ChatGPT Ne de DeepSeek…

As the AI sector becomes increasingly competitive, recent tests have showcased which model stands as the world’s best artificial intelligence. The findings were not particularly surprising.

Google’s relatively new AI model, Gemini 2.5 Pro, achieved results superior to other models during various tests. The performance gap between Gemini 2.5 Pro and its competitors was notably impressive. Although it lagged behind in a few assessments, the score differences were not very significant.

Gemini 2.5 Pro excels in multiple tests!

Gemini 2.5 Pro did not just shine in a single test. In a mathematics-specific test called MathArena, the model demonstrated a vastly superior performance compared to its counterparts. It recorded a score of 24.4%, while AI models from firms like OpenAI, Anthropic, and DeepSeek could not exceed 5%. Additionally, in the Humanity’s Last Exam test, which assesses when AI surpasses human performance in a domain, Gemini 2.5 Pro outperformed its rivals.

However, this does not mean that Gemini 2.5 Pro is the best AI model in every area. In the coding-specific test LiveCodeBench, the model scored 70.4%, while OpenAI’s o3-mini model obtained a score of 74.1%. In another software-related test, SWE-bench, Gemini 2.5 Pro achieved a score of 63.8%, falling behind Claude 3.7 Sonnet, which scored 70.3%.