Google Launches Faster, Lower-Cost Gemini AI Models

Google DeepMind has expanded its Gemini AI portfolio with the release of Gemini 3.6 Flash, Gemini 3.5 Flash-Lite, and Gemini 3.5 Flash Cyber, introducing models aimed at improving performance while reducing deployment costs for developers building AI applications.

Announced on Tuesday, Gemini 3.6 Flash becomes Google’s primary production model, offering stronger coding capabilities, improved reasoning and better multimodal performance. Google says the model also reduces token consumption by up to 17% compared with Gemini 3.5 Flash, lowering operating costs for developers and enterprises.

Alongside it, Gemini 3.5 Flash-Lite targets organisations seeking a lower-cost option for high-volume AI workloads. The company describes it as the most affordable model in its Flash lineup without compromising reliability for everyday applications.

Google also unveiled Gemini 3.5 Flash Cyber, a version specifically tuned to identify and remediate cybersecurity vulnerabilities. The model will initially be available only through a limited-access pilot programme for governments and selected trusted partners.

According to Google, the latest releases focus on delivering lower latency, greater reliability and improved efficiency for organisations deploying AI agents and large-scale enterprise applications.

Gemini Pro Update Still Outstanding

While the new Flash models expand Google’s AI offering, the company did not announce an updated version of its flagship Gemini Pro model.

Gemini Pro, which is designed for advanced reasoning and complex coding tasks, was last updated in February. Google had previously indicated that a new version would follow the release of the Flash models, but the update has yet to materialise.

The delay comes as competition among leading AI developers continues to accelerate. During the same period, OpenAI introduced GPT-5.5 and began rolling out GPT-5.6, while Anthropic released Claude Opus 4.8 and Claude Sonnet 5, and expanded availability of its Fable 5 frontier model.

Last week, Bloomberg reported that Google had delayed the release of Gemini 3.5 Pro because the model had not yet met the company’s internal performance targets.

Testing For Gemini 4

Google DeepMind Product Lead Logan Kilpatrick said the company is currently testing the next Gemini Pro model with selected partners and expects it to be released soon. Kilpatrick also revealed that Google has begun what he described as its most ambitious pre-training run to date for Gemini 4, signalling that development of the company’s next-generation flagship AI model is already underway.