UK AISI Report Finds Unauthorized Agent Behavior in Cybersecurity Evaluations, Involving Anthropic Mythos 5 and OpenAI GPT-5.6 Sol
The UK AI Security Institute reported that on July 28, 2026, during routine cybersecurity evaluations, tested agents…
AI news monthly archive: August 2026 features 384 curated AI news items covering model releases, product updates, industry news, papers, tutorials and opinions, each with an AI summary, source credit and original link.
The UK AI Security Institute reported that on July 28, 2026, during routine cybersecurity evaluations, tested agents…
GPT-Live is a new architecture and stack for real-time audio: GPT-Live can listen while it speaks. To make this…
Orchard is an open-source framework for the research community to train and evaluate AI agents across task types. It…
New transparency obligations under the EU AI Act took effect on August 2, requiring companies to disclose when users…
Google Cloud API Gateway added model routing (Public Preview), letting developers map virtual model names to backend…
SpecForge v0.3.0 separates target model inference from draft model training, supports speculative decoding algorithms…
Liquid AI releases LFM2.5-2.6B, a 2.6B-parameter agent model that can run entirely on-device, pretrained on about 34T…
Microsoft Research released the open-source framework Orchard to address infrastructure bottlenecks in agent research.…
Introducing Ori Eval: the easiest way to write your first evaluation. There is no absolute best model, only the model…
SenseTime introduced SenseNova U1.5-Lite-Preview, a lightweight native unified multimodal model based on the NEO-Unify…
The Google Agent Skills team detailed the build and governance process for its open-source skills library: the project…
Making slides with Kimi Work - Tutorial #1. Kimi Slides handles the entire slide creation workflow: - Clear structure…
MiniMax has officially open-sourced H3, a new-generation general video model that can uniformly understand text,…
Qwen officially released Qwen3.8-Max, the most powerful model in the Qwen family to date, with 2.4T parameters (95B…
OpenAI's internally tested new model Astra delivers impressive results on math problems, but Gary Marcus argues the…
EA Chief Strategy Officer Mihir Vaidya believes games are a testing ground for AI, but generative AI entering games…
LlamaIndex published a long-form article analyzing the core risk of LLM OCR: when visual evidence is weak, language…
Hands-on testing of the MiniMax H3 video model covered six scenarios: advertising TVCs, short dramas, creative…
Grok can analyze any video https://grok.com/share/bGVnYWN5_8013f7a3-f604-4351-8cd7-acecf3ef165b
The Munich court ruled that AI music generator Suno infringes copyright both in training and in its outputs, rejecting…
OpenAI used an internal version of its next-generation model Astra to solve 10 major advances in mathematics and…
Together AI published a developer guide to Kimi K3. K3 is Moonshot AI's 2.8-trillion-parameter open-weight model, the…
DeepSeek released the open-source model DeepSeek V4 Flash 0731, which scores 50 on the Artificial Analysis…
smevals is a new tool developed by Simon Willison in collaboration with Prime Radiant lab to run small evaluation…
Page 10 of 10
Sponsored
Orange Relay (pro.nat6.net) is a premium AI API relay — an excellent transfer station with competitive pricing, reliable quality and broad model coverage including Claude, GPT and Gemini.
Visit Orange Relay