Skip to content

ModelsGoogle DeepMind

Google introduces Gemini 2.5 Pro, its first 2.5 thinking model, debuting at No. 1 on LMArena

Gemini 2.5 Pro at a glance

Gemini 2.5 Pro Experimental reasons before answering, shipped with a 1M-token context window and led several math, science and coding benchmarks at launch.

What changed with Gemini 2.5 Pro

On March 25, 2025, Google introduced Gemini 2.5, calling it its most intelligent model, starting with an experimental version of 2.5 Pro. It is a thinking model that reasons before responding, and Google said it debuted at the top of the LMArena human-preference leaderboard by a significant margin.

Google reported leading scores on GPQA and AIME 2025 without test-time tricks such as majority voting, 18.8% on Humanity's Last Exam without tools, and 63.8% on SWE-Bench Verified with a custom agent setup. It shipped with a 1 million token context window, with 2 million promised, in Google AI Studio and the Gemini app for Advanced users, with Vertex AI and paid pricing to follow.

Gemini 2.5 Pro in numbers

Values are copied from the linked source on the date shown. They are not IA Compa measurements.

Context windowas of · Source for Context window: blog.google (opens in a new tab)
1M tokens

How it compares

Other models in the same category that we have covered, using each model's latest recorded value. (Models)

Context window (tokens)higher is better

Where to compare Gemini 2.5 Pro

Independent leaderboards and official resources for comparing this model yourself. Rankings change often; each site shows its own current data.

Why Gemini 2.5 Pro matters for businesses

For businesses, 2.5 Pro combined strong reasoning with very long context, making it practical to analyze whole contracts, codebases or long recordings in one request, and it reset expectations for what Google's models could do.

Sources for this story

  1. Official
  2. Gemini 2.5 Pro model page (opens in a new tab)

    Artificial Analysis · artificialanalysis.ai

    Benchmark

Last verified: . IA Compa summarizes public announcements for business readers. We are not affiliated with the companies mentioned. Always confirm details, pricing, and availability with the official source.

  • #llm
  • #google
  • #gemini
  • #reasoning
  • #long-context

Want this working in your business?

We turn new AI capabilities into agents, voice receptionists, and automations you can actually use.