Is generative AI technology ready to replace human workers or is it the best tool to help workers? Latest, what Google has done is clearly that this AI technology is the best tool to help humans because the Gemini 2.5 Pro reasoning AI model is so powerful that it can generate code for an entire video game with just a single line of prompts.
According to Google, Gemini 2.5 Pro is their first model to successfully beat OpenAI, Anthropic, xAI, and DeepSeek in benchmark tests. Google's new reasoning AI model is also more powerful and faster in handling mathematical problems, code generation, and can generate not only text but also audio, images, and video. Google labels Gemini 2.5 Pro Experimental as a reasoning model that is easier to understand user commands no matter how complex.
Gemini 2.5 Pro is still in the testing phase and can only be accessed through Google AI Studio. In addition, Gemini Advanced users on the web and apps can also test it but with a context window limited to 1 million tokens. Support for 2 million tokens will be provided in the future. This model will also be available on the Vertex AI platform.