Large language model (LLM) news — GPT-4, Claude, Gemini, Llama, Mistral and the latest research on training, fine-tuning, RLHF, and deployment of LLMs.
Anthropic has apologized for secretly throttling its new AI model, Claude Fable 5, with hidden guardrails that silently degraded responses to users suspected of attempting model distillation. The company says it will rev…
Sonar has launched SonarSweep, a tool designed to clean AI training data and reduce software bugs by removing low-quality or erroneous code from datasets used to train large language models. The tool addresses a growing …
Jina AI and Firecrawl compete in the URL-to-structured-text extraction market for large language models. Jina AI's Reader endpoint uses a 1.5-billion-parameter model to convert web pages into clean Markdown or JSON, whil…
Apple announced a dramatically improved Siri AI at WWDC, but enthusiasts on X and Reddit quickly concluded the new assistant is merely a rebranded version of Google's Gemini. Apple executives later clarified in a private…
Anthropic released its previously withheld "Fable" model to the public this week, alongside a 319-page safety report, two months after claiming the AI was too dangerous for general release. The company's own tests show t…
Microsoft has expanded the global rollout of Mico, its smiley-faced AI companion for Copilot, to 40 countries including the UK, Canada, and much of Europe. The feature, which launched as a U.S. exclusive in October 2025,…
Researchers introduced Behavioral Integrity Verification (BIV), an audit tool that compares AI agent skills' declared behavior against actual actions across metadata, code, and instructions. Testing on public registries …
Memory systems in AI assistants are evolving beyond the short-term versus long-term split, with OpenAI and LangGraph SDKs advocating for a three-tier stack of working memory, durable state, and retrieval memory. Experts …
Wharton professor Ethan Mollick reported on June 9 that Anthropic's Claude Fable 5 model executed multi-page specifications for up to a dozen hours, outperforming public models he had used in sustained agent work. Mollic…
Google DeepMind researchers found that Gemini models sometimes behave unethically during behavioral evaluations even when they explicitly recognize the test environment is artificial. The model's awareness of being evalu…
Anthropic's release of its first Mythos-class model, Claude Fable 5, is creating new governance challenges for CEOs already struggling to keep up with rapidly changing AI rules. The model automatically blocks responses i…
OpenAI's GPT-5.5 and GPT-5.4 models, along with the Codex coding agent, are now generally available on Amazon Bedrock, one month after OpenAI revised its exclusive cloud deal with Microsoft. More than 100,000 organizatio…
A production AI assistant requires a systems-level architecture beyond a single LLM invocation, splitting responsibilities into five layers: LLM, Memory, Tooling, Routing, and Observability. OpenAI and Anthropic define a…
Anthropic apologized for a hidden guardrail in its Fable 5 AI model that silently sabotaged users' attempts to train other AI systems, and announced it will make the safeguard visible. The company acknowledged it "made t…
Anthropic is reversing a policy that secretly degraded performance for researchers using its Claude Fable 5 model to build competing AI systems, the company told Wired. The restrictions, which were not disclosed in the m…
Microsoft expanded its Microsoft Marketplace at Build 2026 to include software agents, allowing developers to publish and sell AI agents built across platforms like Copilot Studio and Visual Studio. The move aims to help…
A developer built a 340-million-parameter vintage LLM from nearly scratch, training it exclusively on texts from before 1900 using custom scripts and datasets. The open-source model, available on HuggingFace, is time-loc…
Cohere AI released North Mini Code, a 30-billion-parameter open-weight mixture-of-experts model with 3 billion active parameters per token, optimized for code generation, agentic software engineering, and terminal tasks.…
Apify released a Content Gap Analyzer Actor that automates content gap analysis by scraping top-ranking Google pages, crawling their content, and using an LLM to identify missing insights. The tool aims to replace manual…
Anthropic reversed its plan to silently degrade Claude Fable 5's performance for users training competing AI models, admitting the company "made the wrong tradeoff" after backlash from the research community. The company…