US government forces Anthropic to disable Fable 5 and Mythos 5 worldwide

US orders Anthropic to suspend Fable 5 and Mythos 5 for all foreign nationals The US Commerce Department issued an export control directive citing national security, forcing Anthropic to abruptly disable its newest models for all customers. Anthropic says the government cited a jailbreak technique for minor vulnerabilities, but argues the action was based on a misunderstanding. 12345678

OpenAI retires GPT-5.2 and migrates all users to GPT-5.5 As of June 12, OpenAI deprecated the GPT-5.2 family in ChatGPT, automatically moving conversations to corresponding GPT-5.5 models. The quiet swap may alter outputs for tuned prompts and workflows, prompting developers to verify behavior. 9

OpenAI faces investigation from coalition of state attorneys general New York's attorney general served OpenAI with a subpoena seeking documents on advertising, user engagement, model sycophancy, data handling, and treatment of minors. The probe follows a separate Florida lawsuit and adds to OpenAI's legal challenges. 10

OpenAI launches GPT-5.4 with million-token memory and multi-step workflows The new model introduces a million-token context window and native multi-step reasoning capabilities, positioning it as a major upgrade. The launch comes amid conflicting reports about the GPT-5.x family lineup. 11

Google unveils DiffusionGemma, a diffusion-based text model with 4x faster inference The experimental open-source model generates entire passages simultaneously rather than token-by-token, achieving up to 4x speedup on local hardware. It is built on Gemma 4 and Gemini Diffusion research. 12

Google sues Chinese crime group for using Gemini AI in mass phishing scams The lawsuit alleges the group exploited Gemini to generate convincing phishing messages at scale. Google seeks to disrupt AI-powered fraud and set a legal precedent for model misuse. 13