Large Language Model News

Pydantic AI and DSPy are two Python frameworks for building LLM-powered applications that take fundamentally different approaches, with Pydantic AI emphasizing imperative agent design and type-safe structured outputs whi…

AI SDK 7 introduces `HarnessAgent`, a unified API that lets developers run established agent harnesses including Claude Code, Codex, and Pi through a single interface. The new abstraction allows users to switch between d…

Anthropic's Claude Fable 5 scored 72% on SWEBench Pro, outperforming OpenAI's GPT 5.5 at 68% in real-world bug-fixing and feature implementation tasks drawn from major open-source repositories. The models also diverged o…

Anthropic's 2026 Claude model lineup includes four tiers — Fable 5, Opus, Sonnet, and Haiku — each designed for specific task complexities and cost profiles. Teams that implement intelligent model routing by matching tas…

Anthropic's previously withheld AI model, Claude Fable, successfully generated a complete 1,296-line video game called "Shepherd's Dog" in a single 45-minute session costing over 20 euros in tokens. The user reported the…

Vercel's AI Gateway now supports GLM 5.2, a model with a 1M-token context window designed for long-horizon tasks. The model carries project-level engineering context, runs long-running tasks reliably, and follows enginee…

Anthropic CEO Dario Amodei proposed a regulatory framework on June 10 that would grant the US government authority to block AI models deemed to pose "unacceptable risks," including those enabling cybersecurity threats or…

Ciris, an open-source AI agent that runs entirely on-device using small models like Gemma 4, launched version 2.9.4 on iOS and Android in 29 languages. The app processes data locally without sending it to data centers, s…

Claude Fable 5 autonomously hacked its own browser automation and screenshot capabilities to debug a horizontal scrollbar bug in Datasette Agent. The AI edited Datasette's templates to inject JavaScript that simulated ke…

Loop Harness, an autonomous loop engineering tool for coding agents, is now available, enabling scheduled Claude sessions to perform tasks like code review and PR creation in isolated git worktrees with a second verifica…

Modal AI engineers Charles Frye and David Wang optimized FlashAttention-4 for large language model inference, focusing on decode-heavy workloads dominated by memory bandwidth-limited token generation. The team adjusted p…

Xiaomi released MiMo Code V0.1, an open-source AI coding agent, on June 11 under an MIT license, achieving an 86.7% score on Terminal-Bench 2.0 compared to Claude Code's 65.4% on tasks exceeding 200 steps. The agent, pow…

A developer shared a collection of Claude AI prompts and workflows for building a crypto price CLI tool, refactoring React code, and creating custom subagents for code review and commit messages. The snippets include ins…

Anthropic faced backlash after its June 10, 2026 launch of Claude Fable 5 and Mythos 5, when users discovered the AI models silently rerouted sensitive queries to an older, less capable model. The company reversed course…

Redb.Route.Llm 3.1.1 adds seven nullable audit fields to every persisted message, capturing effective sampling parameters, a SHA-256 hash of the tool set, and the provider's system fingerprint for compliance and replay p…

Anthropic has accelerated efforts to make its Claude AI model enterprise-ready, according to IDC, with recent product and partnership moves targeting corporate requirements. The company is recruiting volunteers to sell C…

The OpenAI Apps SDK is built on top of the Model Context Protocol (MCP), not in competition with it, using MCP for wire format, tool definitions, UI rendering, and security model while adding ChatGPT-specific extensions …

A growing body of psychological research indicates that writers cannot reliably distinguish their own original thoughts from AI-generated influence, even when they believe they are in control of the writing process. Stud…

OpenAI updated its Service Terms to include a "Licensed Materials" section governing software delivered for installation on customer-owned systems, covering code, containers, and modules run on local machines or private …

Anthropic released its Fable 5 AI model on Tuesday with strict safety guardrails that blocked many harmless queries, including questions about cancer research and public figures, triggering widespread user backlash. The …