Gemini has launched their most intelligent AI model to date, Gemini 2.5. With key developments that centre Gemini 2.5 models as “thinking models” making it capable of reasoning through their thoughts before responding.
Key highlights of Gemini 2.5 Pro Experimental include:
- State-of-the-art performance on a wide range of benchmarks, debuting at #1 on the LMArena leaderboard by a significant margin, indicating a highly capable model with high-quality style.
- Strong reasoning and code capabilities, leading on common coding, maths, and science benchmarks. It has shown significant improvement in coding performance over Gemini 2.0.
- A state-of-the-art score of 18.8% on Humanity’s Last Exam (across models without tool use), a dataset designed to capture the human frontier of knowledge and reasoning.
- Availability now in Google AI Studio and in the Gemini app for Gemini Advanced users, with availability on Vertex AI coming in the following weeks.
- It ships with a 1 million token context window, with 2 million coming soon.
- Thinking capabilities are built in by default (“Thinking On”).
For more details visit Gemini 2.5
