Here is today's AI news. Story 1. OpenAI. Released GPT-6 Astra, a new frontier model aimed at autonomous computer use and multi-step agent workflows, with availability expanding through ChatGPT, the API, and enterprise platforms.. OpenAI has started rolling out GPT-6 Astra, a new frontier model focused on computer use, coding, research, and multi-step agent work. It’s first available to a limited set of organizations, with broader access coming soon through ChatGPT Plus, Pro, Business, Enterprise, the API, and AWS. OpenAI says Astra adds safety monitoring for agent mistakes and can complete complex tasks end to end. Story 2. Google. Launched Gemini 3.8 Flash and Gemini 3.8 Flash Cyber, positioning the new models as stronger at reasoning, coding, and cybersecurity tasks while retaining prior Flash-tier pricing.. Google launched Gemini 3.8 Flash, an upgraded workhorse model for coding, agentic workflows, and multi-step reasoning, alongside Gemini 3.8 Flash Cyber, a restricted cybersecurity variant for vulnerability discovery and automated patching. Google says Flash keeps the same introductory Flash-tier price. The move sharpens competition in enterprise AI and defensive security, with Cyber limited to vetted defenders via the Fairwind Program. Story 3. Meta. Introduced Muse Spark 1.3, an updated AI model focused on coding and agentic tasks that Meta says delivers frontier-level performance at substantially lower cost.. Meta released Muse Spark 1.3 on September 2, positioning it as a major upgrade for coding and agentic work. The model is available in Muse Code and via Meta’s Model API, where it can sustain longer tasks and manage multiple workflows. Meta says it uses about 20% fewer tool calls and 25% fewer tokens than Muse Spark 1.2. It scored 75.4 on DeepSWE v1.1 and 88.8 on Terminal-Bench 2.1, matching GPT-5.6, while improving the economics of AI-agent deployment. Story 4. Anthropic. Released Claude Fable 5.1 and Claude Mythos 5.1, model updates aimed at developers and coding agents, alongside lower cache-read pricing.. Anthropic released Claude Fable 5.1 for general use and Claude Mythos 5.1 for vetted cybersecurity and life-sciences organizations. The models target long-running agentic coding, research and document workflows, with a 1-million-token context window and 128,000-token maximum output. Input and output remain $10 and $50 per million tokens, while cache reads fell 75% to $0.25, potentially lowering costs for repetitive agent workloads. Story 5. U.S. Department of Justice / OpenAI. The Justice Department formally backed OpenAI’s fair-use position in its copyright case with *The New York Times*, its first stated position in the major AI-training copyright disputes.. The Justice Department filed a statement of interest backing OpenAI in The New York Times copyright suit, saying AI training on copyrighted text can qualify as fair use. The DOJ called LLM training “extraordinarily transformative” and warned a contrary ruling could “severely hamper” U.S. AI progress, innovation, and national security. It’s the first federal government merits position in the AI-training copyright fights. That's all for today's AI news briefing.