China's Qwen model now accounts for a growing share of the AI models used by Western companies to serve, train, and build AI, according to a co-written article by Dean Meyer and Konstantine Buhler. Th…
Moonshot released Kimi K3, a 2.8 trillion-parameter open-weight model, the largest ever, with native vision and a 1M-token context window, claiming ~2.5× better scaling efficiency than K2 and pricing …
Zhipu AI's GLM model family spans a 6,000x range in hardware requirements, with the flagship model needing more memory than a typical home has RAM, according to a user who tested GLM-4.7-Flash on a 16…
Apple's on-device generative AI service, Apple Intelligence, has been registered with China's cyberspace regulator for use on iPhones, clearing a key requirement for a mainland rollout that will incor…
Thinking Machines' model Inkling has been added to an arena where 27 frontier AI models, including OpenAI's GPT-5 series, Anthropic's Opus 4.1, and Google's Gemini 3, compete by generating 52 small ap…
A user reports that AI models differ drastically in how long they can run unattended on a host with sudo without causing system issues: Opus breaks hosts within 2 agent-hours, Gemini 3 Pro within 12, …
China's Cyberspace Administration approved Apple Intelligence for on-device generative AI use in the country, clearing the largest regulatory hurdle for Apple's AI suite in its second-biggest market. …
Fireworks AI has reached a $17.5 billion valuation as enterprises increasingly seek cheaper alternatives to proprietary AI models, the company announced. The inference startup, which operates a pay-pe…
Apple Intelligence, Apple's generative AI offering, received approval from China's Cyberspace Administration of China for launch in the country, with integrations of Alibaba's Qwen AI model and Baidu'…
China's first national regulation on emotional interaction with artificial intelligence took effect July 15, 2026, forcing ByteDance and Alibaba to shut down personalized assistant features on their D…
Chinese AI firms have surged to over 30% of AI token usage among US firms weekly since US export controls on Anthropic's Mythos 5 and Fable 5 models in June 2026, with DeepSeek, Z.ai, and Alibaba's Qw…
Chinese users of AI-powered companion bots are bidding farewell to their virtual partners as national regulations took effect Wednesday to curb emotional dependency. Major AI providers including ByteD…
China's Cyberspace Administration issued the Interim Measures for the Administration of AI Anthropomorphic Interactive Services on July 15, 2026, banning virtual intimate-relationship services for min…
China on Wednesday enforced sweeping regulations banning AI companion features, forcing tech giants including ByteDance, Alibaba, and Tencent to remove virtual relationship tools from their platforms.…
VaultCharts has launched a free desktop trading app that integrates AI analysis with traditional charting tools, allowing users to bring their own AI models from providers including OpenAI, Anthropic,…
Apple is partnering with Baidu to power AI search for its Apple Intelligence suite in China, with Baidu developing an AI-powered search experience that enhances Siri and enables image and text underst…
A Thai developer known as 9arm has released a small Claude Code skills repo, 9arm-skills, that introduces a cost-saving technique for AI workflows. The key skill, qwen-agent, delegates repetitive, low…
Simple self-distillation (SSD) improves LLM code generation by fine-tuning models on their own sampled outputs, boosting Qwen3-30B-Instruct from 42.4% to 55.3% pass@1 on LiveCodeBench v6, with gains c…
Chinese regulators have granted Apple a licence to launch its Apple Intelligence AI service in China, with Alibaba and Baidu as technical partners. The Cyberspace Administration of China (CAC) publish…
A new speedrun benchmark for training reinforcement learning agents to solve Sokoban puzzles has been released, with two tracks: an LLM track fine-tuning Qwen3-4B-Instruct-2507 on 8xH100 GPUs to achie…