Google introduces Veo 3 Fast, an optimized model for speed and price, along with new image-to-video capabilities for both Veo 3 and Veo 3 Fast, enabling developers to efficiently create high-quality video content from text or still images, with varying pricing based on the model and audio inclusion, now available in the Gemini API.

LangExtract is a new open-source Python library powered by Gemini models for extracting structured information from unstructured text, offering precise source grounding, reliable structured outputs using controlled generation, optimized long-context extraction, interactive visualization, and flexible LLM backend support.

The Gemini Embedding model enhances AI applications, particularly through context engineering, which is being successfully adopted by various organizations across industries to power context-aware systems, leading to significant improvements in performance, accuracy, and efficiency.

Gemini 2.5 Flash-Lite, previously in preview, is now stable and generally available. This cost-efficient model is ~1.5x faster than 2.0 Flash-Lite and 2.0 Flash, offers high quality, and includes 2.5 family features like a 1 million-token context window and multimodality.

The `logprobs` feature has been officially introduced in the Gemini API on Vertex AI, provides insight into the model's decision-making by showing probability scores for chosen and alternative tokens. This step-by-step guide will walk you through how to enable and interpret this feature and apply it to powerful use cases such as confident classification, dynamic autocomplete, and quantitative RAG evaluation.