Google is launching some new AI models, but not the one many expected.

The company on Tuesday announced three Gemini models that it said are faster and cheaper for running AI agents, as Google continues to double down on raw power efficiency.

This isn’t entirely a choice: Google said the Gemini 3.5 Pro, the new frontier model it originally promised for June, is still in testing. Google also teased the Gemini 4, the next big flagship, although it’s not expected to arrive for several months.

Of the new models released this week, Google describes 3.6 Flash as the “workhorse” that is better than 3.5 Flash at tasks such as coding. Google said 3.6 Flash reduces the use of tokens (AI-processed text units) by up to 17% compared to its predecessor, according to an analysis by Artificial Analysis Index, a benchmark for comparing AI tools.

Google also introduced 3.5 Flash-Lite, which it says is the fastest and “most cost-effective” version of its 3.5 model yet.

Then there’s 3.5 Flash Cyber, which is designed to find and fix cybersecurity vulnerabilities at a similar level of performance as its rivals but at a “lower price per token than larger models,” said Tulsee Doshi, senior director of product management for Google’s Gemini group. he said in a blog post.

That last part is crucial. Anthropic and OpenAI have launched cybersecurity models in recent months, but it appears that Google is once again hoping to undercut the competition. Google said Flash Cyber ​​will initially be available exclusively to governments and trusted partners.

Realizing that using many tokens does not always guarantee the best results, more and more companies are controlling their AI spending and turning to more cost-conscious models. That gives Google a strategic opening.

“Companies are already spending their annual token budgets and we are only in May,” said Google CEO Sundar Pichai he said earlier this year. “If companies used a combination of Flash and other cutting-edge models, they could save a lot of money.”

Google says its new models hit the “sweet spot” of efficiency and power for running agents.

No sign of 3.5 Pro

Gemini 3.5 Pro, Google’s next big frontier model, is being tested with partners and will ship “as soon as it’s ready,” the company said.

Business Insider reported in June that Google had pushed back the planned release date for 3.5 Pro to July. Bloomberg recently reported that the model could have been delayed further.

As of Tuesday, Google did not have a single model in the top 10 in the Artificial Analysis rankings, which uses a composite benchmark to score models in areas including mathematics and reasoning.

Google also said Tuesday that it has begun its “most ambitious pre-training yet” for Gemini 4, its next big model.

Google typically releases these big flagship models late in the year, and it’s still unclear whether the 3.5 Pro will be powerful enough to push Google to the forefront in the meantime.

However, as companies take a more considered approach to their AI spending and limit their use of tokens, Google is positioning itself for the time being.