Google Unveils Latest AI Model Gemini 2.5

Google announced on the 25th (local time) that it will launch the latest artificial intelligence (AI) model Gemini 2.5.

Google introduced Gemini 2.5 as "our most intelligent model" and "a thinking model designed to solve increasingly complex problems." It has been just over three months since the launch of Gemini 2.0 in December last year.

Google explained, "The performance has improved and accuracy enhanced because it can reason through thinking before responding," adding, "Reasoning ability goes beyond simple classification and prediction to analyze information, draw logical conclusions, reflect context and nuance, and make decisions based on information."

After releasing Gemini 2.0 in December last year, Google introduced its first reasoning model, 'Gemini 2.0 Flash Thinking,' in January. Google stated that it plans to directly integrate reasoning capabilities into all models going forward.

Google first released an experimental version of 'Gemini 2.5 Pro' as the first model in the Gemini 2.5 product line. This model is available on Google's developer platform, Google AI Studio, and the subscription service Gemini Advanced (monthly subscription fee $20).

Google claimed that Gemini 2.5 Pro outperformed its previous state-of-the-art AI models and major competitor AI models in several aspects. In tests evaluating code editing, this model scored higher than top AI models from OpenAI, Anthropic, and Chinese AI startup DeepSeek.

In tests measuring software development capabilities, it surpassed OpenAI's reasoning model o3-mini and DeepSeek's R1, and in multimodal evaluations based on math, humanities, and science problems, Google reported that it showed better performance than major competitor models.