Google cuts enterprise AI costs with new Gemini models

Google DeepMind has released three proprietary AI models—Gemini 3.6 Flash, Gemini 3.5 Flash-Lite, and Gemini 3.5 Flash Cyber—designed to reduce the enterprise cost of running autonomous software agents. The release prioritizes operational efficiency over maximum capability, directly targeting the unit economics of corporate AI deployments.

The flagship of the trio, Gemini 3.6 Flash, is priced at $1.50 per million input tokens and $7.50 per million output tokens. The actual savings for enterprises extend beyond this sticker price, as the model requires fewer tokens to complete complex tasks. According to the Artificial Analysis Index, it uses 17% fewer output tokens than its predecessor, with savings reaching 65% on long-horizon software engineering benchmarks like DeepSWE.

For enterprises prioritizing volume and speed, Google introduced Gemini 3.5 Flash-Lite at $0.30 per million input tokens and $2.50 for output. The model processes 350 output tokens per second, roughly twice the speed of the older 3.1 Flash-Lite. Despite its lower cost, it outperforms the standard Gemini 3 Flash on key coding and agentic benchmarks, making it a practical tool for massive document processing and high-volume search tasks.

Google is also targeting the cybersecurity sector with Gemini 3.5 Flash Cyber, a specialized model fine-tuned to identify and fix vulnerabilities. It will integrate directly with Google's CodeMender agent and will be available exclusively to governments and trusted partners. While Google has not set a specific price for the Cyber model, it noted it will cost less per token than larger models.

Notably absent from the release is Gemini 3.5 Pro, the flagship model Google previously indicated would arrive this summer. Rivals OpenAI and Anthropic have already released more powerful flagship updates since Google's prior Pro model debuted in February 2026. Addressing the delay on social media, Google technical staffer Logan Kilpatrick wrote: "Gemini 3.5 Pro is currently testing with partners and we plan to make it broadly available as soon as it’s ready."

The decision to roll out highly efficient mid-tier models first suggests Google sees immediate enterprise demand for cheaper, faster agentic systems. All three new models are closed-source, locking corporate customers into Google's API ecosystem to access these cost savings. Both 3.6 Flash and 3.5 Flash-Lite feature a 1-million-token context window and a knowledge cutoff of March 2026.