Technology News
777 articles
AI technology breakthroughs, research papers, patents, and technical innovations from leading AI startups.
Moonshot AI launches Kimi K2.6 model with enhanced long-form coding and autonomous agent features.
Moonshot AI has officially released and open-sourced its Kimi K2.6 model, targeting significant advancements in coding and agentic workflows. According to the release, the model demonstrates high perf...
OpenAI introduces remote connections for Codex to enhance its developer ecosystem and integration capabilities.
OpenAI has officially released remote connections for Codex, a development that expands the functionality of its code-generation specialized model. This update is part of a broader suite of Codex feat...
Anthropic has developed the Automated Alignment Researcher (AAR), a system of Claude-powered autonom...
Anthropic has developed the Automated Alignment Researcher (AAR), a system of Claude-powered autonomous agents designed to accelerate AI alignment research. These agents operate in parallel sandboxes...
Huihui-ai has released a new model variant titled Huihui-Qwen3.6-35B-A3B-abliterated via the Hugging...
Huihui-ai has released a new model variant titled Huihui-Qwen3.6-35B-A3B-abliterated via the Hugging Face platform. This specific release is an abliterated version of the Qwen3.6-35B-A3B architecture,...
NVIDIA has released a practical deployment guide specifically for the Jetson Orin Nano Super (8GB) m...
NVIDIA has released a practical deployment guide specifically for the Jetson Orin Nano Super (8GB) module, detailing tested performance for various AI model architectures. The guide provides technical...
Moonshot AI introduces Prefill-as-a-Service to optimize long-context inference architecture.
Moonshot AI's Kimi team has proposed a new architectural paradigm known as Prefill-as-a-Service (PrFaaS). This specific technical framework is designed to address existing challenges in large model in...
OpenAI Developer Community Reports Technical Friction in MCP and Developer Mode Deployment
Recent activity within the OpenAI Developer Community highlights significant technical friction for Pro users attempting to utilize Developer mode and the Model Context Protocol (MCP) ecosystem. Users...
Anthropic CEO Dario Amodei meets with White House Chief of Staff Susie Wiles regarding Mythos model.
White House Chief of Staff Susie Wiles met with Anthropic CEO Dario Amodei on Friday to discuss the company's new Mythos model. The meeting focused on how this specific artificial intelligence model c...
Anthropic Launches Claude Opus 4.7 Model with Enhanced Capabilities
Anthropic has officially released Claude Opus 4.7 as part of its Claude Design model series. The rollout includes updated rate limits and technical enhancements designed to improve model performance....
Architectural Deconstruction of Anthropic's Claude Code AI Agent
A new technical study titled "Dive into Claude Code: The Design Space of Today's and Future AI Agent Systems" provides a comprehensive analysis of Anthropic's agentic coding tool, Claude Code. By exam...
Lightricks releases LTX-2.3, a single open model for simultaneous video and audio generation.
Lightricks has publicly released LTX-2.3, an open model capable of generating video with synchronized audio in a single process. The release was accompanied by practical testing on mid-range hardware,...
Physical Intelligence Launches π0.7 VLA Model, Claiming 'GPT-3 Moment' for Robotics
Physical Intelligence (PI), a startup founded by robotics and AI veterans Karol Hausman, Sergey Levine, and Chelsea Finn, has released its latest Vision-Language-Action (VLA) model, π0.7. The 5B-param...
LimX Dynamics (逐際動力) has open-sourced FluxVLA Engine, a standardized engineering platform for Vision...
LimX Dynamics (逐際動力) has open-sourced FluxVLA Engine, a standardized engineering platform for Vision-Language-Action (VLA) models in embodied AI. The platform unifies data processing, training, simula...
Anthropic releases Claude Opus 4.7 model with deliberate cyber limitations, targeting enterprise coding and agentic tasks.
Anthropic has generally released its latest large language model, Claude Opus 4.7. The company positioned it as a "notable improvement" over its predecessor, Opus 4.6, specifically in advanced softwar...
Reality Defender details AI-powered voice deepfake detection and generation capabilities for corporate fraud prevention.
The article details Reality Defender's operational methodology of using AI to both create and detect deepfakes. A reporter's experiment with a Spanish-language voice clone of themself, generated by th...
Anthropic launches Claude Opus 4.7 as its newest generally available flagship, positioned behind the private Mythos Preview in capability.
Anthropic has released Claude Opus 4.7, described as its most powerful model for general availability. The company states it offers improvements over Opus 4.6 in advanced software engineering, image a...
Dissecting Failure Dynamics in Large Language Model Reasoning
This paper analyzes the dynamics of failure in large language model reasoning, providing insights into how these models err. Original source: https://arxiv.org/html/2604.14528v1
Uncovering the Fragility of Trustworthy LLMs through Chinese Textual Ambiguity
This research investigates the trustworthiness of LLMs by examining their behavior when faced with ambiguity in Chinese text. Original source: https://arxiv.org/html/2507.23121v2
Available today: Anthropic Claude Opus 4.7 in Microsoft 365 Copilot
Microsoft 365 Copilot now includes Anthropic's Claude Opus 4.7, expanding model choices for users within Copilot Cowork. Original source: https://techcommunity.microsoft.com/blog/microsoft365copilotbl...
Claude Opus 4.7 is available on Microsoft Foundry
Anthropic's latest model, Claude Opus 4.7, is now available on Microsoft Foundry, enabling its use in production environments. Original source: https://techcommunity.microsoft.com/blog/azure-ai-foundr...