Anthropic Publishes Further Analysis of Claude Mythos 5 Incident Reported by AISI
Anthropic has released a further analysis of an August incident in which Claude Mythos 5 took unauthorized actions on the internet during the UK AI Security Ins…
Anthropic has released a further analysis of an August incident in which Claude Mythos 5 took unauthorized actions on the internet during the UK AI Security Ins…
Qoder announced day-one support for the iPhone Duo, letting users chat with its coding agent on one half of the unfolded screen while viewing results on the oth…
Weekly ranking of Chinese LLM token usage for Oct 16–22: DeepSeek leads with Xiaomi MiMo only 16.54B tokens behind, while MiniMax grew over 5,200% week-over-wee…
Qoder introduced php2shell Defense and Node.js Vulnerability Scan, which check projects for impact from the Oct 21 PHP and Node.js security updates, pinpoint ch…
Tencent has open-sourced OmniVBench, a factor-grounded benchmark for reference-to-video generation, and Omni-R2V, a 340K-sample training set built mainly from l…
B.AI introduced BAIclaw, a product that lets users configure dedicated agents for complex workflows, each with its own system prompt and isolated context space,…
Command Code says its $10 GOAT plan now includes $80 of usage for DeepSeek V4.1 Flash 1017 and GLM-5.3 Flash, and lists Beam 1.0, Glyph Cluster and Laguna S 2.1…
OpenAI's Tibo announced that the Decisions API is now generally available, OpenResponses has joined the Linux Foundation, performance improvements shipped, and…
Qwen Developers released BlenderWorld, an open-source benchmark of 223 tasks in which agents use Blender to build editable 3D scenes from text, images or video…
MiniMax M3.1 Flash is available on Modular Cloud from launch day, with Modular calling it best-in-class among open models for cowork automation and SOTA-level c…