Google has launched Gemini 2.5 Pro, describing it as their “most intelligent AI model” to date. The new model represents a significant advancement in Google’s AI capabilities, with a particular focus on reasoning abilities that are now built directly into the system.
According to Google’s announcement, Gemini 2.5 models are “thinking models” that can reason through their thoughts before responding, leading to enhanced performance and improved accuracy. This reasoning capability allows the model to analyze information, draw logical conclusions, incorporate context and nuance, and make informed decisions.
The experimental version of Gemini 2.5 Pro has achieved impressive benchmark results, claiming the top position on the LMArena leaderboard, which measures human preferences, by a significant margin. Google reports that the model demonstrates strong reasoning and code capabilities, leading on common coding, math, and science benchmarks.
Technical capabilities
The new model builds on previous Gemini features while adding several technical improvements:
- A 1 million token context window (with 2 million coming soon)
- Native multimodality supporting text, audio, images, video and code
- Advanced reasoning for complex problem-solving
- Improved coding performance over Gemini 2.0
On specialized benchmarks, Gemini 2.5 Pro scored 18.8% on Humanity’s Last Exam without tool use, which Google claims is state-of-the-art. For coding tasks, it achieved 63.8% on SWE-Bench Verified with a custom agent setup.
Availability and access
Gemini 2.5 Pro is currently available through:
- Google AI Studio for developers
- The Gemini app for Gemini Advanced subscribers
- Coming soon to Vertex AI
Google plans to introduce pricing in the coming weeks, which will enable higher rate limits for production use. The company also indicates that future models in the Gemini 2.5 family will incorporate these thinking capabilities directly.
Unlike previous versions, Google is no longer explicitly attaching the “Thinking” label to the model, though users can choose to “Show thinking” in the Gemini app to see the AI’s train of thought.
Sources: Google, 9to5Google