Gemini 2.5 Pro: Google ships native reasoning in its frontier multimodal model
In one sentence Google DeepMind ships Gemini 2.5 Pro, first model in the 2.5 family with built-in 'thinking'. 1M context window, reasoning capabilities competitive with o1/o3.
Google makes a significant jump with Gemini 2.5 Pro. For the first time, the model "thinks" before answering — a process similar to OpenAI o1 or Claude with extended thinking. The difference: it's integrated into the model itself, not a separate model.
Results: it takes first place in many independent benchmark leaderboards (LMArena, ai-explained.video). It can read up to 1 million tokens at once — an entire codebase, hours of video, or dozens of documents.
For developers: free in Google AI Studio to try, paid via API. Also available to Gemini Advanced (consumer) users and Vertex AI (enterprise).
Companies
Google DeepMind
Tools
Gemini 2.5 Pro
Tags
Sources