We’re launching Lyria 3.5 in Google Flow Music, with advances across musicality, lyrics, vocals, and creative control
deepmind.google·6d ago
AI Engineering
DeepMind Blog·14d ago·5 min read
Introducing Gemini 3.6 Flash, 3.5 Flash-Lite, and 3.5 Flash Cyber
TL;DR
DeepMind has launched Gemini 3.6 Flash, 3.5 Flash-Lite, and 3.5 Flash Cyber, enhancing AI capabilities for various applications. These models focus on improved performance and efficiency in processing tasks.
✦ Why It Matters
Engineers can integrate Gemini 3.6 Flash into their applications to enhance real-time data processing capabilities.
Key Takeaways
- Utilize 3.6 Flash for cost-effective coding tasks.
- Implement 3.5 Flash-Lite for high-speed document processing.
- Explore 3.5 Flash Cyber for enhanced cybersecurity measures.
- Provide feedback to improve future Gemini models.
Full Summary
Gemini 3.6 Flash improves upon its predecessor by offering 17% better token efficiency and lower costs, priced at $1.50 per million input tokens. It excels in coding and knowledge tasks, achieving higher precision and reduced verbosity in benchmarks.
The 3.5 Flash-Lite model is designed for high-speed tasks, delivering 350 output tokens per second at a competitive price of $0.3 per million input tokens. Additionally, 3.5 Flash Cyber focuses on cybersecurity, efficiently identifying and fixing vulnerabilities through the CodeMender platform.
These models are now available for developers and enterprises, with 3.5 Pro in testing for future release. The advancements aim to streamline agentic workflows and enhance performance across various applications.