Introducing Google's Latest AI Models: Gemini 2.5 Pro and Flash
Google has recently launched two new additions to its Gemini AI model family, Gemini 2.5 Pro and Flash. These models are now accessible for developers, signifying a significant milestone in Google's AI advancements.
Gemini 2.5 Pro: Advancing AI Reasoning Capabilities
Gemini 2.5 is built upon Google DeepMind's hybrid-reasoning architecture, aiming to provide robust AI reasoning capabilities with enhanced efficiency and speed across various applications. This latest version is a substantial improvement over its predecessors, positioning Google competitively against OpenAI and their renowned GPT models. Offering enhanced creative, conversational, and coding abilities, Gemini 2.5 Pro includes a unique "configurable thinking budget" feature, enabling developers to balance speed and computational resources effectively. This enhancement addresses previous challenges in long-form coherence and formatting, making the model more dependable for production purposes.
Gemini 2.5 Flash-Lite: A Cost-Efficient AI Model
Google has also unveiled Gemini 2.5 Flash-Lite, an experimental AI model known for its speed and affordability within the 2.5 family. Despite being in the preview phase, this model retains essential features such as multimodal reasoning, tool access, and a reasoning "thinking" mode, while significantly reducing latency and costs. The model is priced at around $0.1 per million input tokens and $0.4 per million output tokens, making it a cost-effective option compared to Gemini 2.5 Pro, which costs $1.25 per million input tokens and $10 per million output tokens.
Integration with Google Search
With these new updates, Gemini 2.5 models are now widely available across Google AI Studio, Vertex AI, and the consumer Gemini app. Custom variants of these models already play essential roles in Google Search, strengthening Gemini's standing in the competitive AI landscape alongside OpenAI, Anthropic, and others. Google continues to expand its "thinking" model portfolio, offering developers and users a diverse range of AI solutions.
