LG Research has introduced EXAONE 4.5, an open multimodal language model capable of simultaneously analyzing text and images.
Google has added two new request processing modes to the Gemini API – Flex and Priority – allowing developers to choose between speed and cost.
AI: Events
Google Gemma and NVIDIA: Powerful AI Right on Your Computer
Technical context • Infrastructure
Google has released a new family of Gemma models, optimized in collaboration with NVIDIA for local execution – from compact edge devices to powerful workstations.
Google has released Gemma 4 – the company's most powerful open models to date, focused on complex reasoning and agentic scenarios.
Google has released the Gemma 4 family of open models, and AMD has provided immediate support on release day across its entire hardware spectrum, from data centers to laptops.
Red Hat and NVIDIA have jointly achieved leading results in the independent MLPerf Inference v6.0 test, which covers image recognition, speech, and reasoning tasks.
AI: Events
Gemma 4: Google DeepMind's Multimodal AI That Runs Directly On-Device
Technical context • Products
Google DeepMind has released Gemma 4 – an open family of multimodal models that process text, images, video, and audio directly on-device.
AI: Events
Qwen3.6-Plus: Alibaba's New Model on the Path to True AI Agents
Technical context • Products
Alibaba has released Qwen3.6-Plus, an updated multimodal model with enhanced agent capabilities, a one-million-token context, and improved code support.
AI: Events
When One GPU Isn't Enough, and a Second Is Too Costly: A New Approach to Running AI in Production
Infrastructure
Two new open-source projects offer a way to run multiple AI models on a single GPU with dynamic memory management, without sacrificing performance.