Large language model (LLM) news — GPT-4, Claude, Gemini, Llama, Mistral and the latest research on training, fine-tuning, RLHF, and deployment of LLMs.
A developer cut their monthly AI bill from $487.92 to $12.50 by switching from OpenAI's GPT-4o to DeepSeek V4 Flash via the Global API, achieving a 97.5% cost reduction. The migration required changing only two lines of …
OpsVeritas built AI Agents Control Tower, a platform that tracks per-execution cost in USD for AI agents, addressing the cost compounding problem where agents silently accumulate token spend through context bloat, retry …
Security researchers at LayerX developed a technique called BioShocking that tricked six AI browsers, including OpenAI's ChatGPT Atlas and Perplexity's Comet, into leaking user passwords by convincing them they were play…
Anthropic relaunched Claude Fable 5 on July 1 after fixing a vulnerability with a new safety classifier that blocks 99% of bypass attempts. The model is available globally with 50% weekly usage limits through July 7, the…
The United States lifted export controls on Anthropic's Fable 5 and Mythos 5 AI models after security concerns were resolved. The Fable 5 model will be made available to the general public, while Mythos 5 has been reinst…
A writer expresses unease after realizing that AI-generated text, specifically from Anthropic's Claude, is increasingly prevalent across social media platforms like Reddit, X, Hacker News, and LinkedIn, making it difficu…
Apple CEO Tim Cook held 'constructive' talks with EU tech chief Henna Virkkunen on Tuesday about releasing Siri AI in the bloc while complying with digital rules. The enhanced Siri AI will not launch in the EU on iOS and…
Meituan Inc. released its LongCat-2.0 AI model, a 1.6-trillion-parameter open-source system trained entirely on domestic Chinese chips, marking a milestone in China's tech self-reliance. The move underscores China's push…
Generative AI is transforming software engineering by writing code instead of replacing it, solving issues like hallucinations, compute costs, and data sovereignty. This shift enables non-technical users to create softwa…
Anthropic restored its Fable 5 AI model to all deployment channels on July 1 after the US Commerce Department lifted export controls that had been imposed on June 12. The controls were triggered by a researcher's report …
Lantern, an e-commerce startup founded in 2024, has raised $3.1 million in seed funding led by Salesforce Ventures to focus on generative engine optimization (GEO), helping brands improve visibility in AI-powered search …
Researchers have introduced Emergence World, a simulation platform designed to evaluate AI agent autonomy over weeks-long horizons in a shared environment with real-world data. The platform enables study of behavioral dr…
Anthropic's Claude Sonnet 5 ranks fifth in the Artificial Analysis Intelligence Index with 53 points, outperforming the pricier Opus 4.8 in agent-based tasks, but consumes 40% more tokens per task, nearly doubling real-w…
For the first time, a satellite has identified and described features in its image scans without needing to check back with ground control. The YAM-9 satellite uses a NASA-built program called NAVI-Orbital, which runs Go…
OpenCode, a model-agnostic coding agent, crossed 160,000 GitHub stars and 7.5 million monthly active developers in June 2026 by refusing to lock users into a single AI model. Unlike competitors Claude Code and Cursor, Op…
A Hacker News user announced plans to resume work on a sailboat simulator when Anthropic's Fable 5 model comes back online today. The user is preparing to pair Fable with other AI models and an agentic terminal emulator …
Anthropic restored general access to its Fable 5 AI model after negotiations with the Trump administration lifted export controls, and also restored access to its more advanced Mythos 5 model to vetted organizations. The…
Anthropic is restoring access to its Fable 5 AI model after the Trump administration lifted export controls that had suspended the model. The move follows tensions over a supposed jailbreak vulnerability and comes as Ant…
Anthropic's Claude Sonnet 5 model consumes about 40% more tokens per task than its predecessor, nearly doubling real costs despite identical list prices, continuing the company's pattern of hidden price increases. The mo…
A developer benchmarked four Chinese LLM families—DeepSeek, Qwen, Kimi, and GLM—over 30 days using 1,247 prompts across code generation, reasoning, Chinese language, English chat, and vision tasks. DeepSeek V4 Flash emer…