AI News : Deepseek Returns, Amazons Secret AI Models, Googles Breakthrough , Veo 3 Beaten and More

Deep Seek released a new open source AI model, Deepseek version 3.2, that achieved gold level performance in various contests like the international mathematical olympiad. This model pushes the frontier of AI from textbook problems to competitive creative problem solving.

Mistral released Mistral Large 3, a flagship model in their family of experts. This model, although not as good as its Chinese counterparts, is Europe's only open-source model and competes in the AI space.

Anthropic's Opus 4.5 showed dominance in benchmarks like humanity's last exam and Arc AGI 2, surprising many with their performance and progress in different areas like coding.

Nano Banana Pro is a breakthrough model capable of understanding and generating complex AI tasks with hyperrealistic outputs, pushing towards achieving Artificial General Intelligence (AGI).

An analysis of various video models like VO, Sora, Runway Gen 4.5 XL, and Cling in generating realistic images and videos, showcasing their strengths and weaknesses in different scenarios.

Anthropic showcased remarkable progress in robotics, with a robot capable of performing complex movements and tasks, demonstrating the potential for advancements in the field.

The EU faces pressure and criticism in regulating AI, with debates on data anonymization, privacy laws, and rules for high-risk AI systems, highlighting the challenges in balancing competitiveness and regulation in the AI race.

Google introduced Titans and Mirs, new AI architectures mimicking human brain functions with a focus on unexpected information importance and memory updating during operation, setting a potential paradigm for future AI models.