OpenAI has launched GPT-6 Astra, its most capable artificial intelligence model to date. The model can operate computers, complete multistep workflows and handle tasks that would otherwise require continuous human input.
By Chris Borden Meta has released an artificial intelligence (AI) model called Muse Spark, marking its “first step” toward a comprehensive overhaul in this field.
By Daniel Mercer Chinese startup Zhipu AI has released its flagship language model GLM-5.2 for long agentic tasks and programming. The open-source solution has a 1 million-token context window, an MIT license, and support for local deployment.
By Alex Rowland On June 9, Anthropic released two versions from the same Claude model family. The company describes Fable 5 as a Mythos-class solution that is safe for general use. Claude Mythos 5 is a “private” base model with reduced restrictions in selected areas.
By Daniel Mercer OpenAI has announced several changes to ChatGPT. The recently released GPT-5.5 Instant is getting an update designed to make responses feel more natural, easier to read, and better structured, with fewer long bullet-point lists.
By Chris Borden Alibaba's Qwen team has introduced Qwen3.7-Max — a proprietary model designed specifically for agentic tasks. In a practical test, the model ran fully autonomously for 35 hours optimizing a hardware kernel, producing results that benchmark it alongside Anthropic's Claude Opus 4.6.
By Chris Borden Google has unveiled Gemini 3.5 Flash — what the company is calling its most powerful AI model for coding and agent creation. The model can write code autonomously, manage research projects, and build an operating system from the ground up.
By Daniel Mercer Cursor has released Composer 2.5, a significantly improved AI coding model for its in-house code editor. The model is based on Moonshot’s open-source Kimi K2.5 checkpoint and was trained on 25 times more synthetic training tasks than its predecessor, Composer 2. According to Cursor, 85% of the compute was used for additional training and reinforcement learning.
By Chris Borden OpenAI has introduced GPT-5.5-Cyber — a specialised variant of its flagship AI model designed for cybersecurity tasks, currently available to a limited group of developers.
By Chris Borden Baidu has introduced Ernie 5.1, a new language model that builds on the pretraining foundation of Ernie 5.0 while using a much smaller architecture.
ES 
EN