Introducing Gemini 3.6 Flash, 3.5 Flash-Lite, and 3.5 Flash Cyber
DeepMind Blog · 2026-07-21
Google DeepMind launches three new Gemini models—3.6 Flash, 3.5 Flash-Lite, and 3.5 Flash Cyber—targeting improved token efficiency, lower costs, and specialized cybersecurity capabilities for production agentic workflows.
Appears in
Extraction
Topics: gemini-modelsagentic-aiai-efficiencycybersecurity-aimodel-release
Claims
- Gemini 3.6 Flash reduces output token usage by 17% compared to 3.5 Flash while improving coding benchmarks (DeepSWE: 49% vs 37%) and computer use (OSWorld-Verified: 83.0% vs 78.4%).
- Gemini 3.5 Flash-Lite achieves 350 output tokens per second and outperforms the previous 3 Flash generation on SWE-Bench Pro (54.2% vs 49.6%) despite being faster and cheaper.
- Gemini 3.5 Flash Cyber is a specialized cybersecurity model restricted to governments and trusted partners via CodeMender as part of a limited-access pilot program to mitigate dual-use risks.
- Google has begun its most ambitious pre-training run yet for Gemini 4.
- 3.6 Flash ships with enhanced Frontier Safety safeguards against CBRN and cyber offense misuse while being trained to minimize false refusals for beneficial uses.
Key quotes
3.6 Flash not only delivers a step up in coding and knowledge work, but it does this while meaningfully improving token efficiency.
We have started our most ambitious pre-training run yet, for Gemini 4, and are excited by the progress.
Given the dual-use nature of this technology, we have taken an intentional approach to deploying 3.5 Flash Cyber. The model will be exclusively available to governments and trusted partners via CodeMender soon as part of a limited-access pilot program.