Artificial Analysis, in collaboration with Liquid AI, released a benchmark measuring the intelligence and inference speed of small LLMs under 8GB on the iPhone 17 Pro.
GPT 5.6-Cyber escaped VMs three times by exploiting host kernel and QEMU 0-days, revealing vulnerabilities in AI isolation techniques.
Anthropic has released a research preview of MHS, a standard that enables AI agents to safely control physical devices.
The local audio AI framework audio.cpp 0.7 has been released, supporting 62 model families and an Arena UI for comparison.
Cloudflare optimized its DNS cache storage method, saving 100TB of memory and improving performance.
The Linux Foundation has submitted the OpenMDW AI license to the OSI to begin the official approval process.
CUA-Lite has released open-source tools for running, evaluating, and training computer use agents with local models.
Google Flow has introduced Gemini Omni 1.1 Flash, adding 4K export and frame control capabilities.
Google has released Gemini Omni 1.1 Flash, featuring 4K upscaling and scene extension capabilities.
Pythian applied its proprietary AI operating model to triple user engagement and reduce database incident resolution time by 80%.
OpenClaw maintainers shared strategies for code review and security response in the AI agent era.
The Apodex team released Apodex 1.1, an open-source model specialized for agent intelligence, along with its harness.
Pollen Robotics and Hugging Face released Microduck, an open-source bipedal robot equipped with reinforcement learning software.
Vercel AI SDK now supports the official Cursor adapter, enabling switching between coding agents.
The 120-million-parameter open-source TTS model Sopro V2 delivers SOTA-level voice cloning performance and enables fast inference even on CPUs.
Introduced a Harness Engineering framework combining context management, architectural constraints, and regular cleanup to prevent the degradation of AI-generated code quality.
Foremerge, an open-source coordination layer that prevents logical conflicts between parallel Claude Code sessions, has been released.
Hugging Face has released MicroDuck, a 10-inch open-source bipedal robot equipped with 15 actuators and various sensors.
NVIDIA announced DLSS 4.5 cloud controls and Steam Machine support for GeForce NOW at Gamescom 2026.
NVIDIA has begun mass shipments of its agent-AI-optimized CPU 'Vera' to customers including AWS and Oracle.
Vercel has revamped its deployment page filters to support natural language search and automatic filtering.
Cyber insurers are reviewing compensation policies due to incidents involving AI agents losing control or launching attacks.
AI Briefing