Google has unexpectedly made its newest AI model, Gemini 2.5 Pro, available to all users for free, not just premium subscribers. The model has quickly gained recognition for its exceptional reasoning abilities and is currently leading industry benchmarks.
Gemini 2.5 Pro stands out for its massive 1-million token context window (with plans to expand to 2 million), allowing it to process entire code repositories or multiple long documents in a single prompt. The model also features a generous 64,000-token output limit, significantly higher than the approximate 8,000 tokens offered by other Gemini models.
Impressive reasoning capabilities
Multiple journalists and developers have praised the model’s reasoning abilities. Ben Dickson from VentureBeat highlighted its multimodal reasoning, describing how the model successfully extracted information from an article to create an SVG graphic and then improved it when shown a rendered version with errors.
Software engineer Simon Willison reported using Gemini 2.5 Pro to implement a new feature across his website codebase. The model identified necessary changes across 18 different files, completing the entire project in approximately 45 minutes.
Transparent thinking process
A key differentiator for Gemini 2.5 Pro is its transparent reasoning process. The model shows detailed step-by-step thinking, presenting ideas in numbered steps with sub-bullets and internal logic that helps users understand how it arrives at conclusions.
According to Matt Marshall from VentureBeat, this structured chain of thought creates “a breakthrough for trust and steerability” as enterprise users can see how the model reached an answer and then validate, correct, or redirect it more confidently.
Surprising free availability
In a move that surprised many industry observers, Google announced that it is rolling out Gemini 2.5 Pro (experimental) to all Gemini app users, not just Gemini Advanced subscribers who pay $19.99 monthly for Google One AI Premium.
The model is already widely available on gemini.google.com and coming soon to Android and iOS apps. Google stated it wants “to get [its] most intelligent model into more people’s hands asap,” though it remains unclear whether the stable version will also be free when it exits the experimental phase.
Gemini 2.5 Pro currently tops the LMArena leaderboard and leads on math and science benchmarks. Despite its somewhat muted launch, which coincided with OpenAI’s viral image generator, Gemini 2.5 Pro appears to be positioning Google as a serious contender in the enterprise AI market.
Sources: VentureBeat, VentureBeat, 9to5Google