Google Introduces Gemini 3.6 Flash and 3.5 Flash-Lite to Scale AI Agents
Published:
Advancing the Flash Series for Agentic Workflows
Google has officially unveiled its latest Gemini models: Gemini 3.6 Flash and Gemini 3.5 Flash-Lite. Designed to provide high efficiency, low latency, and reliability, these models are tailored to empower developers and enterprises building AI agents at scale.
In addition, Google announced Gemini 3.5 Flash Cyber, a specialized security model paired with the CodeMender code security agent.
Key Model Highlights
1. Gemini 3.6 Flash
- Improved Capabilities: Delivers significant upgrades in coding, knowledge work, and multimodal reasoning compared to 3.5 Flash.
- Token Efficiency: Consumes 17% fewer output tokens on average, reducing execution loops and overall costs.
- Pricing: $1.50 per 1M input tokens and $7.50 per 1M output tokens.
- Built-in Computer Use: Computer interaction tools are now natively integrated into the Gemini API and Gemini Enterprise.
2. Gemini 3.5 Flash-Lite
- Ultra-Fast Speed: Generates 350 output tokens per second, making it the fastest model in the 3.5 family.
- Cost-Effective: Priced at $0.30 per 1M input tokens and $2.50 per 1M output tokens.
- Best For: High-throughput production traffic, document processing, and multi-agent workflows.
3. Gemini 3.5 Flash Cyber & CodeMender
- Fine-tuned specifically to detect, validate, and patch software security vulnerabilities efficiently.
- Will be accessible via a limited-access pilot program for governments and trusted partners.
Future Outlook and Gemini 4
Google confirmed that Gemini 3.5 Pro is currently undergoing testing with partners and will be broadly released once ready. Meanwhile, pre-training for the next-generation Gemini 4 model is already underway.