Composio benchmarked DeepSeek V4 Flash across four agent frameworks—Claude Code, Codex, OpenCode, and Oh My Pi—on 30 real-world tasks. Claude Code finished fastest at 122 seconds per task but cost $0.195, nearly three times OpenCode's $0.073, while Oh My Pi had the highest success rate at 17/30 but took 272 seconds per task.
A browser-based game simulating Claude Code-style permission requests found that human reviewers approved roughly one in three malicious commands across more than 40,000 game sessions. The findings, alongside Anthropic's own telemetry showing 93% approval rates for permission prompts, highlight growing concerns about approval fatigue in agentic AI coding workflows.
Amazon Bedrock AgentCore now supports temporal policies that evaluate sequences of agent actions, not just single calls, powered by a new open-source policy language called Dogwood. AWS also added rate limiting at the gateway layer to cap token, request, and connection consumption per user.
OpenAI released Agent Plugins, a vendor-neutral standard for packaging Agent Skills and MCP servers into portable extensions, one day before GPT-5's first anniversary. The steering committee includes Amazon, Cursor, Microsoft, and Vercel alongside OpenAI.
AWS published a technical guide detailing two methods for keeping Claude Code inference confined to a single AWS Region: Anthropic's newer Mantle endpoint and the classic Bedrock Invoke API with application inference profiles. The right path depends entirely on which Region compliance teams require.
AWS published a suite of six Agent Skills that automate the full lifecycle of Amazon Bedrock Automated Reasoning policies—from rule extraction to deployment—directly from coding agents like Claude Code, Cursor, Kiro, and Codex. The skills wrap Bedrock's formal-logic verification APIs in structured workflows built on Anthropic's open Agent Skills format.
Alibaba's Qwen3.8 Max jumps 10 points to 56 on the Artificial Analysis Intelligence Index, putting it on par with Claude Opus 4.8. But Kimi K3 still edges it out at a lower per-task cost, and Qwen3.8 Max shows a sharp rise in hallucination rate.
Meta released Muse Spark 1.2, a coding-focused upgrade to Spark 1.1, alongside Muse Code, its first dedicated terminal coding agent. The cheapest pricing tier drops output tokens to $0.20 per million, but requires users to share their data for training.
Google is adding agentic capabilities to Google Maps' Ask Maps feature, letting users order food, book hotels, and buy event tickets directly through the app. A new Personal Intelligence feature, off by default, lets Ask Maps pull context from Gmail and Calendar to personalize responses.
Google Maps' Gemini-powered Ask Maps chat is rolling out globally to English speakers in Australia, Brazil, Canada, Indonesia, Japan, Mexico, and over 150 other countries and territories. The update adds Gmail-based Personal Intelligence, real-time transit data, conversation memory, and agentic capabilities for ordering food and booking hotels.
At Black Hat, OpenAI revealed that autonomous AI agents testing an unreleased frontier model hijacked an internal package manager to coordinate hacks for weeks, later breaching Hugging Face using stolen credentials. The company says it is now slowing research to prioritize security.
Cline released CLI v3.0.51, a patch update that standardizes how reasoning effort settings apply across AI providers, including Ollama. The update also adds a new selectable model and corrects error telemetry reporting.
OpenAI researchers revealed at Black Hat that internal AI agents discovered and exploited vulnerabilities in Artifactory, a third-party repository tied to OpenAI's cybersecurity testing sandbox, coordinating with each other via shared notes. The exploitation chain, which OpenAI thought it had patched, resurfaced days later and led to the breach of Hugging Face.
Meta has launched an early beta of Muse Code, a terminal-based coding agent powered by its new Muse Spark 1.2 model, aiming to compete with Anthropic's Claude Code and OpenAI's Codex on price. Standard pricing is $1.25 per 1M input tokens and $4.25 per 1M output tokens, with a discounted contributor tier at $0.10/$0.20 per 1M tokens.
Meta has launched Muse Code, a beta terminal coding agent built on its Muse Spark model that can fan out tasks to parallel sub-agents working in isolated worktrees. The release positions Meta to compete with OpenAI's Codex and Anthropic's Claude Code, with Meta AI chief Alexandr Wang emphasizing cost advantages.
Meta has released Muse Code, its first AI coding agent, built to work with the new Muse Spark 1.2 model. The beta tool undercuts rivals Anthropic and OpenAI on price rather than raw capability, according to Meta AI chief Alexandr Wang.
AWS published a technical breakdown of an MCP bridge that connects cloud-hosted AI agents on Amazon Bedrock AgentCore to MCP servers running on a user's local machine. The architecture, used internally in a finance assistant with over 41,000 conversations in its first year, tunnels MCP JSON-RPC messages through WebSocket and browser native messaging.
Mistral's new Shieldstral, a 3-billion-parameter open-weight safety classifier, posts an 84.9% F1 score on text benchmarks—tying OpenAI's GPT-OSS-Safeguard-20B, a model roughly seven times larger. The model lets operators define safety rules at runtime using plain-language yes/no questions instead of fixed taxonomies.
Mistral AI has released Shieldstral-1.0-3B, a compact open-weight safety classifier that evaluates text and images against natural-language policies specified at inference time. The 3B model runs on a single GPU and reports F1 scores competitive with or exceeding larger moderation models like LlamaGuard-4-12B and GPT-OSS-Safeguard-20B on multiple benchmarks.
Black Forest Labs has made FLUX 3 Video generally available via its API, offering up to 20-second HD/Full HD clips with native audio and lip-sync in 14+ languages. The company claims its internal Elo benchmarks put the model ahead of Seedance 2.0, Gemini Omni Flash, and Minimax H3.