Gemini 2.5 Pro Expands AI Access and Pricing

Gemini 2.5 Pro is now available in a public preview phase, allowing broader access to Google’s innovative artificial intelligence model. This advancement features increased rate limits through the Gemini API and Google AI Studio, further solidifying Google’s commitment to enhancing their AI capabilities to attract developers with more adaptable usage options and competitive pricing.

The transition to public preview follows a successful limited-access launch last month, where Gemini 2.5 Pro was introduced as an experimental version with strict usage limitations. Given the surging demand, Google has now elevated these restrictions and introduced a structured two-tier pricing system for developers.

Understanding Gemini 2.5 Pro’s Two-Tier Pricing

Under the new pricing model, developers utilising up to 200,000 tokens each month will incur the following charges:

• $1.25 (approximately ₹107) for every million input tokens
• $10 (approximately ₹858) for every million output tokens

For developers whose usage exceeds the 200,000-token limit, the fees will escalate to:

• $2.50 (approximately ₹214) for each million input tokens
• $15 (approximately ₹1,290) for each million output tokens

The token system encompasses various types of inputs, such as text, images, audio, and video, with the output pricing reflecting the model’s advanced reasoning capabilities. While the experimental version is still accessible free of charge with limited usage, the public preview allows for more comprehensive engagement with the AI.

Competitive Pricing in the AI Market

Google emphasised that its pricing is significantly lower than that of its competitors. For comparison:

• Anthropic’s Claude 3.7 Sonnet charges $3 per million input tokens and $15 per million output tokens.
• OpenAI’s GPT-4 model starts at $15 per million input tokens and can go up to $60 per million output tokens for outputs, although discounted rates are offered for cached tokens.

The public preview of Gemini 2.5 Pro forms a part of Google’s ongoing initiative to strengthen its position within the rapidly growing generative AI sector. The model has not yet been integrated into Vertex AI, but Google confirmed that its inclusion is forthcoming.

Dubbed as the “most intelligent AI model to date” by the company, Gemini 2.5 Pro supports a variety of multimodal inputs and is meticulously designed for complex reasoning tasks, positioning itself as a direct competitor to the top offerings from OpenAI and Anthropic.