Models & Labs

Google DeepMind Launches Gemini 3.6 Flash Models

Google DeepMindJuly 21, 2026high confidence

Why it matters

→The new models enhance token efficiency, reducing costs for AI workflows.
→3.5 Flash Cyber offers advanced capabilities in cybersecurity, addressing vulnerabilities efficiently.
→These releases demonstrate Google's commitment to advancing AI agent technology.

Google DeepMind Launches Gemini 3.6 Flash Models — ©Google DeepMind

Google DeepMind has introduced new models in its Gemini series, including the 3.6 Flash, 3.5 Flash-Lite, and 3.5 Flash Cyber. The 3.6 Flash model improves token efficiency by 17% compared to its predecessor, offering better performance in coding and knowledge work. The 3.5 Flash-Lite model is optimized for speed and cost, delivering 350 output tokens per second. The 3.5 Flash Cyber model, designed for cybersecurity, will be available to governments and trusted partners. These models are now accessible through various platforms, including Google AI Studio and the Gemini app.

Read original

More from Google DeepMind

Market & Regulationbusiness

Google DeepMind commits $40M to Genesis Mission

Google DeepMind is making a substantial investment in the future of scientific research by allocating $40 million in AI tokens and cloud credits to the Genesis Mission. This initiative is designed to accelerate the pace of American scientific discovery by providing advanced AI tools to researchers at the Department of Energy's National Laboratories. Tools such as AlphaEvolve and AlphaFold 3 are already being utilized to expedite complex mathematical and biological research. This commitment not only enhances the capabilities of scientists but also highlights the transformative potential of AI in driving scientific breakthroughs.

Google DeepMindJul 22, 2026

More in Models & Labs

Models & Labsmodels

llama.cpp b10083 Release Expands Platform Support

The latest b10083 release of llama.cpp continues its trend of broadening platform compatibility, making it a versatile choice for developers across different systems. Notably, this update includes support for Ubuntu with ROCm 7.2, enhancing performance for AMD GPU users. Windows users benefit from updated CUDA support, with DLLs for both CUDA 12.4 and 13.3, ensuring compatibility with the latest NVIDIA technologies. While no groundbreaking new features are introduced, the release solidifies llama.cpp's position as a flexible inference runtime across diverse hardware setups.

llama.cpp ReleasesJul 23, 2026

Models & Labsmodels

llama.cpp b10085 release updates Qwen3-VL interpolation

The latest b10085 release of llama.cpp addresses a key issue with the Qwen3-VL vision model's position embedding interpolation. By aligning the interpolation method with the transformers reference, the update ensures more accurate grounding coordinates, particularly for larger and non-square images. This change is crucial for developers working with image processing tasks, as it reduces discrepancies in image scaling. While the update doesn't introduce new models, it enhances the precision of existing functionalities, making llama.cpp a more reliable tool for AI developers.

llama.cpp ReleasesJul 23, 2026

Models & Labsmodels

Llama.cpp b10089 Release Enhances CUDA Support

The latest b10089 release of llama.cpp significantly enhances CUDA support by adding k-quant and i-quant capabilities to the GET_ROWS function. This update allows for more efficient device-side embedding lookups, reducing the need to fallback to the host and improving performance for single-device graphs. By factoring out super-block dequantizers into shared device functions, the release ensures that all quantized GGML types can now take the direct device path. This marks a notable improvement in CUDA's handling of quantized data, making it more robust and efficient.

llama.cpp ReleasesJul 23, 2026