Google launches new Gemini model trio, teases Gemini 4

Google launched Gemini 3.6 Flash, 3.5 Flash-Lite, 3.5 Flash Cyber and focused on efficiency and pricing, but the biggest item may be that the company teased Gemini 4.

The trio of models landed as all the major LLM players including open source Chinese alternatives have been ramping releases. In comparison to the steady cadence of news nuggets from Anthropic and OpenAI, Google has almost looked quiet in comparison.

In a blog post that rode shotgun along with the launch of its CodeMender security agent, Google outlined the following:

Gemini 3.6 Flash, which is designed to be more efficient than 3.5 Flash. Google said Gemini 3.6 Flash is built on the feedback of customers and developers. The model is designed to be more token efficient across tasks by taking fewer reasoning steps and tool calls for multi-step workflows. In addition, Google said Flash 3.6 is less expensive than its predecessor and is priced at $1.50 per 1 million input tokens and $7.50 per 1 million output tokens.

Note how Google is riffing about efficiency and prices. You can thank open source models and enterprises complaining about AI budgets for that focus.

Gemini Flash-Lite 3.5, which is designed for low-latency and high throughput tasks. Think about agentic search and document processing. Gemini 3.5 Flash-Lite is designed for speed and is priced at 30 cents per 1 million input tokens and $2.50 per 1 million output tokens. Google said the new model "significantly outperforms 3.1 Flash-Lite."

Gemini 3.5 Flash Cyber, which is designed to find and fix security vulnerabilities. The model is built on 3.5 Flash and fine tuned to find, validate and patch vulnerabilities. Gemini 3.5 Flash Cyber is the model powering CodeMender.

Google noted that Gemini 3.5 Pro is currently testing with partners. The company also said “our team is already focusing on building the next generation of models. We have started our most ambitious pre-training run yet, for Gemini 4, and are excited by the progress.”

More:

• Moonshot AI launches Kimi K3
• Hugging Face defends agentic AI attack with Z.ai's GLM 5.2
• Is a FINRA-ish group the answer for frontier AI standards?
• Rightsizing open models may cut your AI inference spend
• SpaceXAI, Meta puts pricing squeeze on Anthropic, OpenAI