Search AI Products and News

Israeli AI coding startup Qodo announced a strategic partnership with Google Cloud aiming to enhance the quality and integrity of AI-generated software. As enterprises increasingly rely on AI to generate large codebases, the demand for efficient oversight and quality assurance tools has never been more urgent. Itamar Friedman, CEO of Qodo, emphasized that AI-generated code is no longer just an auxiliary tool but the foundation of modern development. Imagine

In the rapidly developing field of artificial intelligence today, Alibaba officially open-sourced its new Qwen3 Embedding model series on June 6th. Based on the Qwen3 foundation model, this model is optimized for tasks such as text representation, retrieval, and ranking, marking another significant breakthrough for Alibaba in the AI field. Compared to the previous version, Qwen3 Embedding has achieved over a 40% performance improvement in core tasks such as text retrieval, clustering, and classification. This progress not only makes it more competitive

Recently, the official Qwen announced the release of the Qwen3-Embedding series models, as new members of the Qwen model family. These models are specifically designed for text representation, retrieval, and ranking tasks. The Qwen3-Embedding series is trained based on the Qwen3 base model, fully inheriting its significant advantages in multi-language text understanding capabilities. These models have performed excellently in various benchmark tests, especially showcasing outstanding performance in text representation and ranking tasks. The test used MTE

Camera manufacturers are responding to the rise of AI-generated images by embedding digital signatures to verify real photos. Nikon plans to offer a mirrorless camera with certification technology and a Verify web tool to check the authenticity of images, seeking methods to identify AI-generated content.

Tsinghua Yao Class alumni Ma Tengyu founded Voyage, with the embedding model surpassing OpenAI on the MTEB dataset, becoming SOTA. Voyage's embedding model emphasizes high retrieval accuracy, utilizing self-supervised loss functions and multi-domain training data. Voyage offers both standard and lightweight versions, priced at $0.4 per 1 million tokens, with new users able to try it for free up to 5000 times. This embedding model employs self-supervised loss functions and multi-domain training data, supporting a maximum window length of 4096 tokens.