Google Unveils Gemini 2.5, Outperforming DeepSeek R1, OpenAI o3-mini, and Others in Benchmark Tests
Google Introduces Gemini 2.5: A New Era of AI Models
Earlier today, DeepSeek unveiled its latest large language model, DeepSeek-V3-0324. In response, Google has launched Gemini 2.5, a suite of advanced thinking models touted as their most intelligent to date. As Google describes it, Gemini 2.5 is engineered to handle increasingly sophisticated challenges.
What is Gemini 2.5?
Gemini 2.5 consists of various thinking models, with the Gemini 2.5 Pro Experimental being the standout. This model is reported to outperform its predecessors in key benchmarks, offering remarkable reasoning and programming capabilities.
Features of Gemini 2.5 Pro Experimental
The Gemini 2.5 Pro Experimental model is currently available through Google AI Studio and the Gemini app for users subscribed to Gemini Advanced. Google is also planning to integrate this model with Vertex AI, although no specific timeline has been provided.
Performance Benchmarks
Recent benchmarks reveal that Gemini 2.5 Pro Experimental excels in mathematical tasks and code generation benchmarks, such as AIME 2025 and LiveCodeBench v5. Google states that this version marks significant improvements in coding efficiency compared to Gemini 2.0, enhancing its ability to create and edit web applications and perform complex coding tasks.
Technical Specifications of Gemini 2.5 Pro
Here’s a detailed look at the capabilities and requirements for the Gemini 2.5 Pro Experimental model:
| Specification | Details |
|---|---|
| Model Deployment Status | Experimental |
| Supported Input Data Types | Text, Image, Video, Audio |
| Supported Output Data Types | Text |
| Maximum Input Tokens | 1 million |
| Maximum Output Tokens | 64,000 |
| Knowledge Cutoff | January 2025 |
| Tool Capabilities | – Function calling |
| – Structured output | |
| – Search functions | |
| – Code execution | |
| Best Suited For | – Advanced reasoning |
| – Coding tasks | |
| – Handling complex prompts | |
| Availability | – Google AI Studio |
| – Gemini API | |
| – Gemini App |
Future Developments and Features
Currently, Google has not disclosed pricing for the Gemini 2.5 Pro Experimental model, but updates are anticipated shortly. In recent months, Google has also introduced additional models built on the Gemini framework, including Gemini Robotics—designed specifically for robotic applications—and enhanced features in the Gemini App like Deep Research, which utilizes AI to assist users with research tasks.
With these advancements, Google is poised to redefine the capabilities of AI in both consumer and enterprise applications, pushing boundaries in comprehension, functionality, and user interaction.
