Google Triples the Price of Its Budget AI Model Gemini Flash

Google used its annual I/O developer conference to launch Gemini 3.5 Flash, the latest version of its lightweight artificial intelligence model. While the new model delivers impressive performance gains, its pricing has caught many developers off guard.

A Sharp Price Increase

Gemini 3.5 Flash costs 1.50 dollars per million input tokens and 9 dollars per million output tokens. That makes it roughly three times more expensive than Gemini 3 Flash, which charged just 0.50 dollars and 3 dollars respectively. For developers who relied on the Flash tier as a budget-friendly option for building applications powered by large language models, or LLMs, the jump is significant. Many will need to recalculate their cost projections as they migrate to the new model.

In an unusual move, Google skipped the preview phase entirely and launched 3.5 Flash directly into general availability. The model is already accessible through the Gemini API, Google AI Studio, the Gemini app, and AI Mode in Google Search.

Performance That Justifies the Cost

Despite the sticker shock, the numbers behind Gemini 3.5 Flash tell a compelling story. According to Google, the model outperforms the previous Gemini 3.1 Pro on most coding and agentic benchmarks. It also runs approximately four times faster than comparable frontier models, which are the most advanced AI systems available today. At 1.50 dollars per million input tokens, it remains roughly 40 percent cheaper than the Pro tier at 2.50 dollars.

The emphasis on agents, meaning AI systems that can take actions autonomously rather than just generating text, reflects a broader industry shift. Google is clearly betting that developers will pay more for a model designed to power AI agents that can browse the web, write code, and complete multi-step tasks on their own.

For the broader AI community, the launch sends a clear message. The era of rock-bottom pricing for capable AI models may be ending as companies invest in more sophisticated and powerful capabilities. Developers building token-heavy applications should carefully evaluate whether the meaningful performance gains offset the higher costs before making the switch.