Full-text search across 25231 articles. Combine with topic and date filters; results sorted by relevance.
Chinese lab Moonshot AI released Kimi K3, the largest open-weight model ever, at $15 per million output tokens — roughly a third of Anthropic's Claude Fable 5 — during the World AI Conference in Shanghai last Thursday, t…
Moonshot AI's 2.8 trillion parameter Kimi K3 model, released July 16, scored 1,679 Elo on LMArena's Frontend Code Arena benchmark, beating Anthropic's Claude Fable 5 (1,631) and OpenAI's GPT-5.6 Sol (1,618), and demand w…
The Trump administration is reportedly considering measures to deter adoption of Chinese AI models through sanctions and soft pressure, rather than an outright ban, aiming to protect the market positions of OpenAI, Googl…
China's open-weight AI model Kimi K3, developed by Beijing-based start-up Moonshot AI, has narrowed the country's artificial intelligence gap with the US to 'a matter of weeks', according to analysts. The 2.8 trillion-pa…
Robinhood launched Agentic Trading, a feature that lets users connect third-party AI agents like Anthropic's Claude and OpenAI's ChatGPT to dedicated brokerage accounts for autonomous market analysis, portfolio rebalanci…
Anthropic's Model Context Protocol (MCP), adopted by OpenAI, Google, and Microsoft through 2025 and donated to a Linux Foundation body, has become the standard for connecting AI agents to external tools. A curated list o…
Singapore stocks fell on Monday, with the Straits Times Index (STI) down 0.2% to 5,498.95, amid a tech rout in Asian markets triggered by the launch of Chinese startup Moonshot AI's Kimi 3.0 model, which is comparable to…
Alibaba previewed Qwen3.8, a 2.4-trillion-parameter multimodal AI model, at the World Artificial Intelligence Conference in Shanghai, claiming it is second only to Anthropic's Fable 5, but provided no benchmark scores, m…
China's Moonshot AI and Alibaba released new AI models that they claim rival top US systems from OpenAI and Anthropic at a fraction of the cost, intensifying pressure on Silicon Valley and narrowing America's lead in AI …
Despite a broader tech backlash against 'tokenmaxxing'—spending excessively on AI compute—some companies are doubling down on premium frontier models. Shopify requires engineers to use only top-tier models like OpenAI's …
AI chatbots are more than twice as likely to refuse to generate critical material about leaders from repressive regimes, according to a Meta Oversight Board study released Thursday. The study tested 10 commercial large l…
A new study finds that steering vectors in large language models can be interpreted by training neologisms—new tokens with learned embeddings—that allow the model to describe its own internal concepts. Researchers at Ant…
A new study from arXiv finds that query-aware prompt compression, which invalidates prefix caches on every call, can be more expensive than naive caching under realistic hit rates on Anthropic's Sonnet 4.6 API. The resea…
A new blog post argues that existing methods for decoding internal representations in large language models (LLMs), such as Logit Lens, Tuned Lens, Patchscopes, and Anthropic's Jacobian Lens, are all heuristic and lack a…
The European Parliament is developing an internal AI platform, EPGenAI Hub, to give lawmakers and staff access to models from Meta, OpenAI, Anthropic and Mistral, aiming to provide a safer alternative to public AI tools …
Alibaba's Qwen team previewed Qwen3.8-Max-Preview, a 2.4-trillion-parameter multimodal model with a 1-million-token context window, at WAIC in Shanghai, claiming it ranks second only to Anthropic's Claude Fable 5. Indepe…
Alibaba previewed Qwen 3.8, an open-weight model with over one trillion parameters that handles images, video, documents and text, claiming it is second only to Anthropic's Fable 5. The Qwen team says it outperforms Qwen…
Google, a $4.2 trillion company, has failed to release its flagship Gemini 3.5 Pro model by the promised June 2024 deadline, now delayed to late July, raising concerns about its ability to compete with smaller rivals lik…
Alibaba's Qwen team previewed Qwen3.8, a 2.4 trillion-parameter AI model, on July 19th, claiming it trails only Anthropic's Claude Fable 5 among frontier systems, but withheld technical details such as active parameter c…
The Trump administration launched the Gold Eagle Initiative, a cybersecurity hub to coordinate the detection and patching of vulnerabilities flagged by advanced AI systems, via an executive order from President Trump. De…