Mistral AI has released Voxtral Transcribe 2, a family of speech-to-text models designed for both batch processing and real-time transcription. The company positions the technology as more accurate and significantly cheaper than competing services while enabling on-device processing for sensitive data. The release includes two models. Voxtral Mini Transcribe V2 handles pre-recorded audio files at …
An open-source AI assistant called OpenClaw (formerly Clawdbot and Moltbot) demonstrates how future digital helpers might operate entirely on users’ own computers while offering unprecedented customization and control. Federico Viticci writes on MacStories that the project allows users to interact with powerful language models through familiar messaging apps like Telegram or iMessage. The assistant runs …
Black Forest Labs has introduced a new series of AI image generators called FLUX.2 [klein]. This model family focuses on high speed and low hardware requirements. The company states that these models unify image generation and editing within a single architecture. According to the announcement, the systems produce images in less than one second on …
The open source community has a powerful new tool for high precision image generation. GLM Image, a model from the startup Z.ai, demonstrates superior accuracy in rendering complex text compared to Google’s Nano Banana Pro. Carl Franzen reports for VentureBeat that this new model excels at creating information dense visuals such as infographics and technical …
OpenAI and Google have introduced new tools and models to advance artificial intelligence translation. OpenAI launched ChatGPT Translate, a standalone tool designed to compete with established services. It supports over 50 languages and uses a layout similar to existing translation platforms. Features of ChatGPT Translate A primary feature of ChatGPT Translate is its integration with …
Alibaba’s Qwen team has released Qwen-Image-2512. This update improves their foundation model for generating images from text. The model is now available for public use and enterprise integration. According to the developers, this version focuses on three main areas: human realism, natural detail, and text rendering. The model aims to reduce the artificial appearance often …
The Allen Institute for AI has released Molmo 2, an open-source video model designed to compete with larger proprietary systems in video understanding and analysis. The launch aims to demonstrate that smaller models can serve as practical alternatives for businesses. Emilia David reports for VentureBeat. Ai2 released three versions of Molmo 2: an 8B model …
Researchers at the University of Zurich and Cologne University are creating large language models trained solely on historical documents up to specific dates. The team, led by Daniel Göttlich, Dominik Loibner, Guohui Jiang, and Hans-Joachim Voth, describes their project on GitHub. The models, called Ranke-4B, contain four billion parameters and are trained on 80 billion …
Essential AI has launched Rnj-1, an open-source large language model designed to excel at coding, mathematical reasoning, and scientific tasks. The model, named after Indian mathematician Srinivasa Ramanujan and pronounced “range-1,” marks the company’s first major contribution to the open-source AI ecosystem. The team reports on the Essential AI blog that the 8 billion parameter …
Mistral AI has launched Mistral 3, a collection of 10 open-source AI models designed to run on devices ranging from smartphones to enterprise cloud systems. The French startup released all models under the Apache 2.0 license, allowing unrestricted commercial use. The release includes Mistral Large 3, the company’s flagship model, and the Ministral 3 series …