OpenAI has launched textGrain, an invisible statistical text watermark, as opt-in for API customers worldwide on select models. ChatGPT and Codex output in the EU will be watermarked in the coming weeks, in response to the EU AI Act. OpenAI says detection drops from about 92% to 17% when 25% of words in a 400-token passage are replaced.
OpenAI will add an invisible watermark to text generated by ChatGPT and Codex in the European Union to comply with the EU AI Act's transparency rules. Developers anywhere can enable it on select API models starting today, but it is off by default. OpenAI's own tests show detection falling from about 92% to 66% after 10% of words are replaced with synonyms.
Reflection AI has unveiled Beam, a text-only mixture-of-experts model with 501 billion total parameters, 23 billion active, and a 1 million token context window. The company claims it matches Z.ai's GLM-5.2 on advanced reasoning benchmarks while using 3-4x less inference compute. Weights and the full technical report are due later this month.
Reka AI has released a research preview of Rho-1, a 19-billion-parameter omni-model that processes and generates text, images, video, and robot control actions in a single network. Reka says it uses no tool calls or external models. Context window, pricing, and benchmark scores have not been disclosed.
OpenAI will switch on invisible text watermarks called textGrain for ChatGPT and Codex users in the EU over the coming weeks. API watermarking will be opt-in worldwide, unlike Anthropic's mandatory approach for Claude. OpenAI's own data shows detection drops sharply when text is edited.
Claude Opus 5.5 and Claude Sonnet 5.5 are available on Amazon Bedrock in AWS GovCloud (US) Regions. AWS published a setup guide for running Anthropic's Claude Code against them for regulated workloads, including ITAR. Pricing, context window and benchmark figures were not disclosed.
Amazon SageMaker AI has released the aws-ai-ml skill, distributed through the Agent Toolkit for AWS. It lets MCP-compatible coding agents such as Kiro, Claude Code and Codex benchmark endpoints, recommend deployment configurations and generate executable SageMaker Python SDK v3 code.
GitHub has launched ReviewBench, an open benchmark for evaluating AI code review agents. According to GitHub, it uses representative pull requests, multi-source ground truth, calibrated evaluation, and production-aligned metrics. Dataset size, scores, and licensing were not included in the announcement excerpt.
GitHub has launched ReviewBench, an open benchmark for evaluating AI code review agents. According to GitHub, it uses representative pull requests, multi-source ground truth, calibrated evaluation, and production-aligned metrics. Dataset size, model scores, and licensing details were not included in the announcement summary.
AWS has published the Quick Resource Migrator, a sample MCP server on Amazon Bedrock AgentCore that promotes Amazon Quick resources between AWS accounts in a single tool call. It is idempotent, never deletes from the target, and writes versioned S3 backups before every update.
Axios reports, citing unnamed sources, that Nvidia-backed startup Reflection will soon release its first open-weight model. It is expected to lag the most advanced US models while competing with top Chinese open-weight systems. Parameter count, context window, pricing and benchmarks have not been disclosed.
Google Cloud AI Research and several universities introduced RRSI, a method that stops self-optimizing agent harnesses from memorizing their test tasks. According to the paper, it gains up to 4.7 points on five unseen benchmarks and uses about 30% fewer runtime tokens than the unregularized version, with the underlying model frozen.
An Aleph Alpha study of 967 politically sensitive prompts found that only 17 to 41 percent of responses from Alibaba's Qwen, DeepSeek and Moonshot's Kimi were balanced, according to the company's own AI scorer. DeepSeek V4 Pro refused roughly two-thirds of questions. Aleph Alpha sells "sovereign AI" to governments, which gives it a commercial interest in the result.
Starting in October 2026, Google will limit personal-account Gemini users without a subscription to the smallest model, Flash-Lite. The $4.99/month AI Plus plan loses Pro access, and all three models now require AI Pro ($19.99/month) or higher.
Microsoft and Hugging Face released ThinkingBox, a benchmark that grades AI agents on the database state they leave behind rather than their responses. Across 507 workflows run 20 times each, Claude Opus 5.5 leads at 67.16% pass@1 but passes all 20 attempts on only 241 tasks.
Starting October 9, Gemini app users without a Google AI subscription will be limited to Flash-Lite, losing access to Flash and Pro. AI Plus ($4.99/month) subscribers lose Pro, while AI Pro ($19.99/month) gains Deep Think, according to an updated Google support document.
China Telecom AI's Xing4.0-29B-A4B (formerly the TeleChat line) is a mixture-of-experts model with 29B total and 4B active parameters and a native 256K context window, extensible to 512K. The company claims it is the first model of this scale trained entirely on Ascend NPUs with MindSpore. Community GGUF quantizations from Venastine-Research are already available.
Google-affiliated researchers Benjamin Bratton, Blaise Agüera y Arcas and James Manyika propose "Artificial Symbiotic Intelligence," a framework in which AGI emerges from a social system of people and AI agents rather than a single self-improving machine. The essay, written for the Deepmind Institute, is a conceptual argument and reports no benchmark results.
Anthropic released Mods for Claude Code, a plugin system built on JavaScript and TypeScript functions that hook into events such as tool calls, user prompts and UI rendering. Mods are not sandboxed and run with the user's permissions. They work in the CLI, the desktop app and, partly, the VS Code extension.
Meta introduced Muse Gadgets on Friday, an open-source project with firmware and a Linux SDK for building hardware that connects to its Muse personal AI agent. Meta also built a first-party device, Muse Home Link, and says it is giving away 5,000 units to Muse subscribers.