Skip to content
IFAI

AI models

Google rolls out Gemini 2.5 Pro to free users

Published Reach 50

Google just brought its experimental Gemini 2.5 Pro model to free web users, rolling it out uncharacteristically fast after its paid debut. The model features visible step-by-step reasoning and tops the LMSYS Chatbot Arena leaderboard, though free users miss out on file uploads and get a reduced context window.

Google rolls out Gemini 2.5 Pro to free users
Image: arstechnica.com

Google released its top-tier Gemini 2.5 model last week, initially locking it behind the paid Gemini Advanced paywall. In a rare display of speed, Google has now opened Gemini 2.5 Pro (Experimental) to free users on the web, with mobile app access coming soon.

Thinking mode open to everyone

This 2.5 branch is set to replace Gemini 2.0, which only launched in late 2024. The pace of replacement is remarkably aggressive even by current AI standards.

The headlining upgrade is simulated reasoning. When you prompt the model, you can watch the thinking process unfold in real time. This deliberate step-by-step reasoning helps avoid the common hallucinations and logical slip-ups that plagued earlier iterations.

The results speak for themselves. Gemini 2.5 quickly climbed to the top of the LMSYS Chatbot Arena leaderboard, beating out rival frontier models.

The catches for non-paying users

Naturally, free access comes with a few deliberate roadblocks.

Non-paying users still cannot upload files directly into Gemini. Without file uploads, it is much harder to take practical advantage of the model's analytical brain for tasks like document breakdowns.

You also will not get the full 1 million-token context window available to paying subscribers. Google confirmed the free version has a lower token ceiling, though it hasn't shared the exact cap. Quick testing showed it handles a few thousand words without hiccups, but massive workloads remain off-limits.

Why it matters

Giving free users immediate access to a chart-topping frontier model is a bold flex. AI companies usually reserve their flagship reasoning models for $20-a-month subscribers and hand free users a stripped-down mini model.

We think this creates massive immediate value for casual users and prompt tinkerers. If your workflow consists of pure text generation, brainstorming, or logic puzzles, there is little reason to pay for a subscription today.

However, Google's monetization strategy is clearly visible. By gating document uploads and capping context, Google turns the free model into an irresistible test drive. Once a user actually wants to process real-world work like spreadsheets or lengthy reports, they hit a wall and have to upgrade anyway.

Sources (1)

  1. [1]
    arstechnica.com(opens in a new tab)Originalarstechnica.com

Topic trackerGoogle Gemini model family

Every story on this topic, newest first.

  1. Google debuts Gemini 2.5 Flash with customizable thinking budgets
  2. Google opens Veo 2 video generation to developers
  3. Google brings Veo 2 and Live API to developers, teases Gemini 2.5 Flash
  4. Google brings Deep Research to Gemini 2.5 Pro
  5. Google rolls out Gemini 2.5 Pro to free usersThis story

Best of AI, daily