DeepSeek

Anthropic and OpenAI warn that China is stealing their cutting-edge AI through a technique called distillation, enabling cheap, open-source chatbots that threaten US national security and its lead in …

Indian companies are increasingly adopting Chinese large language models from DeepSeek, Alibaba and Moonshot AI to cut AI costs, deepening India's reliance on China for advanced technology despite bil…

Chinese President Xi Jinping will deliver the keynote at the 2026 World Artificial Intelligence Conference in Shanghai from 17 to 20 July, marking the first time China's leader has appeared at the cou…

Chinese President Xi Jinping will deliver a keynote speech at the opening ceremony of the World AI Conference in Shanghai starting Friday, marking his first appearance at the country's flagship AI eve…

China's Z.ai released GLM-5.2, a free open-source AI model with a 1 million-token context window, drawing comparisons to DeepSeek. In tests, the model performed well on email writing, product recommen…

Open-weight models surged to 29% of AI Gateway token volume in June 2026, up from 11% in April, on under 4% of spend, according to Vercel's AI Gateway Production Index. DeepSeek reached 22.6% of token…

U.S. companies are adopting Chinese AI models such as DeepSeek and Z.ai to cut costs, with usage exceeding 30% weekly since early 2026, according to a Financial Times report. Market participants see t…

An engineer reveals five structural leaks that cause LLM bills to be 40-65% higher than pricing-page estimates, including workload ratio, tokenizer variance, prompt caching, batch processing, and retr…

Google has delayed the launch of Gemini 3.5 Pro from June to July 17, 2026, after Business Insider reported the company is gathering feedback from early users on agent performance and token consumptio…

An engineer who burned $48,000 on a direct AI provider contract built a model router that slashed costs by 97.5% by switching to aggregator-based access. Processing 5 billion tokens per month, the sys…

AMD's open-source GEAK agent-driven framework automated the optimization of the DeepSeekV4 MLA kernel, achieving a 2.10x improvement in end-to-end throughput and a 3.71x reduction in time-to-first-tok…

A new benchmark testing AI models on pull request security reviews found that OpenAI's GPT-5.6 Sol and xAI's Grok 4.5 outperform Anthropic's models, with GPT-5.6 Sol achieving 100% recall at $0.70 per…

A developer released the first open-source training kernels for MiniMax Sparse Attention (MSA) on Hopper and Blackwell GPUs, enabling efficient million-token training with sparse attention. The kernel…

Open-source AI models face a potential ban or severe restrictions within six months as the White House considers new executive orders targeting frontier open-weights models, driven by fears of Chinese…

An engineer audited a $250M-valued AI-agent platform and found that token economics, gated data, and bought distribution prevent autonomous AI from running a company today. The platform spends 61% of …

An engineer at a company built an internal benchmark to evaluate coding LLMs for production use, testing ten models on real-world tasks like Python services and JavaScript debugging. The results showe…

Mesh LLM, a new open-source tool, launched yesterday and hit the Hacker News front page, enabling users to pool GPUs across machines and run AI models up to 235B parameters via an OpenAI-compatible AP…

Goldman Sachs initiated coverage on Chinese AI companies, naming Zhipu and DeepSeek among top picks, and projecting Chinese AI model revenue will grow 25x to $121 billion by 2030. Zhipu's founder defe…

Goldman Sachs released a report on Friday analyzing China's AI large language model industry, finding that Chinese models achieve near parity with US rivals at 10-25% of the cost. The bank projects Ch…

A developer reported that running GPT-4o for a side project led to a $400 monthly AI bill, prompting a shift to Chinese models like DeepSeek V4 Flash and Qwen3-32B, which claim up to 40x cost savings.…