#public-interest

OpenAI opens the Codex harness to developers through Agents API

OpenAI has put the Codex harness behind a public-beta Agents API, adding managed sessions, context compaction, tool coordination, subagents,…

Anthropic launches Claude Fable 5.1 with lower token costs

Anthropic has launched Claude Fable 5.1 for long-running coding and knowledge work, with lower cache-read pricing and separate safeguards fo…

Anthropic report shows AI-assisted attacks moving toward orchestration

Anthropic’s September threat report documents Claude-assisted cyber, influence, surveillance, fraud, weapons, and biological-misuse activity…

OpenAI says GPT-6 Astra helps Devin test and prove its code

OpenAI says GPT-6 Astra is helping Cognition’s Devin test software, return simulator recordings, and document what passed or remains unteste…

OpenAI launches ChatGPT for Financial Services with GPT-6 Astra

OpenAI has launched ChatGPT for Financial Services, an enterprise workspace that combines GPT-6 Astra with built-in financial datasets and s…

IBM releases PatchTST-FM-r2 for open zero-shot forecasting

IBM's Granite Time Series PatchTST-FM-r2 brings open weights, probabilistic forecasts, and a commercial-friendly dual license to zero-shot t…

OpenAI claims an AI-generated solution to Navier–Stokes

OpenAI says an internal model produced a finite-time singularity proof for the Navier–Stokes Millennium Prize problem, with a Lean formaliza…

Cloudflare moves Next.js RCE defenses into baseline WAF rules

Cloudflare has promoted two Next.js remote-code-execution detections from beta into baseline managed WAF signatures. The change strengthens …

GPT-6 Astra reaches GitHub Copilot for long-horizon coding

GitHub has made OpenAI’s GPT-6 Astra generally available in Copilot, with gradual rollout across paid plans and usage-based provider pricing…

Google DeepMind Opens AlphaGenome Atlas to 9 Billion DNA Variants

Google DeepMind has released AlphaGenome Atlas, a free academic portal that precomputes molecular-effect predictions for roughly 9 billion p…

OpenAI expands journalism support from classrooms to newsrooms

OpenAI is extending its journalism program with more than 400 ChatGPT Edu subscriptions for students and faculty at CUNY Newmark and Northwe…

OpenAI Maps Its GPT-6 Strategy Around Delegated Work

OpenAI says GPT-6 Astra, broader distribution, and in-house compute are converging around a larger market for delegated work.

Google puts Lyria 3.5 into the Gemini API for full-song generation

Google's Lyria 3.5 music model is moving from a product showcase into the Gemini API public preview, with full-song generation, vocals, lyri…

xAI retires Grok Imagine Image Quality and changes default billing

xAI will retire the grok-imagine-image-quality API slug on November 2, 2026, routing requests to Grok Imagine Image 2.0 at low quality while…

GitHub CLI Linux package key expires September 5

GitHub's current PGP key for Linux package repositories expires on September 5, 2026. Older apt, yum, and dnf installations should refresh t…

Claude Fable 5.1 reaches GitHub Copilot with retention caveats

OpenAI commits $1B to expand Daybreak cyber defense access

OpenAI is committing $1 billion over six months to subsidize Daybreak access, training, technical support, and partnerships for resource-con…

GitHub Copilot adds Gemini 3.8 Flash with usage-based billing

GitHub is rolling Google’s Gemini 3.8 Flash into Copilot for Pro, Pro+, Max, Business, and Enterprise users, with introductory provider pric…

OpenAI launches GPT-6 Astra with staged plan and API access

OpenAI is rolling out GPT-6 Astra to Trusted Access enterprises first, followed by Plus, Pro, Business, Enterprise and API access in the com…

GitHub reopens Copilot Business signups with upfront billing

GitHub is gradually reopening Copilot Business and Enterprise signups for card and PayPal customers, while changing seat billing and overage…

Vercel and OpenRouter cut GLM-5.3 API pricing for a limited window

Vercel and OpenRouter are offering limited-time 50% pricing windows for Z.ai's GLM-5.3 family, with different model slugs, providers, and en…

GitHub Copilot approvals turn code review into a merge-control decision

GitHub is previewing opt-in Copilot approvals that can satisfy repository merge requirements, with controls at enterprise, organization, and…

Qwen3.8 Max 0902 reaches AI Gateway with a pinned snapshot

Alibaba's Qwen3.8 Max 0902 is now available through Vercel AI Gateway, giving developers a dated model ID for repeatable coding and agent wo…

OpenAI’s Astra raises the bar for cyber-critical model access

OpenAI says its upcoming Astra model is the first to meet its Critical cybersecurity capability threshold, with its strongest cyber abilitie…

Anthropic releases Claude Fable 5.1 with cheaper agentic work

Anthropic says Claude Fable 5.1 is generally available for coding and knowledge work, with lower cache-read pricing and more targeted cybers…

Google releases Gemini 3.8 Flash for long-horizon agent work

Google says Gemini 3.8 Flash improves coding, agentic tasks, and multi-step reasoning at the same introductory price as 3.7 Flash, while Ope…

Qwen-Drive 1.0 puts 3D perception and motion planning in one 4B model

Qwen has published Qwen-Drive-1.0-4B on Hugging Face, combining driving-scene understanding, 3D perception, and motion planning around a Qwe…

CrowdStrike launches SafeMind models for autonomous cyber defense

CrowdStrike introduced SafeMind, a red-team and blue-team model family built with NVIDIA Nemotron for Falcon-based cyber defense.

Claude Code 2.1.257 adds Fable 5.1 and tighter agent safeguards

Claude Code 2.1.257 makes Fable 5.1 the default model and adds stricter controls around metadata access, network paths, credentials, MCP ser…

Google brings agentic video understanding to Gemini Flash

Google has launched agentic video understanding for Gemini 3.7 Flash, 3.6 Flash, and 3.5 Flash-Lite, with claimed gains of up to 88% fewer t…

OpenClaw 2.0 Rebuilds Onboarding, Browser Work, and Shared Sessions

OpenClaw 2.0 is the project's largest update so far, combining simpler setup, a rebuilt browser experience, shared cloud sessions, and broad…

Runway extends legacy Unlimited access through November 30

Eligible legacy Runway Unlimited subscriptions receive three additional months before moving to the credit-based Max plan.

Next.js 15.5.24 and 16.3.3 close two critical RCE paths

Next.js shipped 15.5.24 and 16.3.3 after disclosing a critical Windows-hosted RCE and a separate AVIF image-optimization RCE path.

OpenAI retires the official DALL·E GPT on August 30

OpenAI is retiring the official DALL·E GPT in ChatGPT today and advises users to download images they want to preserve.

Cloudflare brings Z.ai's GLM-5.3 to Workers AI

Cloudflare now serves Z.ai's GLM-5.3 through Workers AI, adding a 1M-token agentic coding model with paid-only access and OpenAI-compatible …

OpenAI launches Rosalind Workbench for reviewable life-science research

OpenAI's Rosalind Workbench brings GPT-Rosalind, scientific viewers, and guided sequencing workflows into a research preview for eligible li…

OpenAI sets November 12 cutoff for models in SpaceX-owned Cursor

OpenAI says it will wind down the contract supplying models to Cursor after SpaceX's acquisition, with a proposed November 12, 2026 shutoff …

Qwen3.8-27B pairs open weights with multimodal agent controls

Qwen3.8-27B is a self-hostable 27B vision-language model with a 262K native context window, adjustable reasoning depth, and Apache 2.0 weigh…

Gemini Omni 1.1 Flash reaches GA with 40-second video extension

Google's production release adds longer scene extension, endpoint-frame control, draft rendering, and upscaled 4K output to Gemini's video A…

Google replaces Gemini Notebook’s daily quotas with compute-based limits

Gemini Notebook will adopt five-hour and weekly compute-based limits for consumer accounts on September 2, replacing predictable daily featu…

Ant Group releases Ling 3.0 Flash Fin for long-context finance work

Ling 3.0 Flash Fin brings a 256K context window and tool calling to financial research workflows, with free API access through September 25.

Anthropic previews Model Hardware Standard for AI-controlled lab equipment

Anthropic’s limited MHS preview gives model-agnostic agents a common interface for microscopes, liquid handlers, robot arms, and other progr…

OpenAI opens commercial operations in Brazil as ChatGPT use doubles

OpenAI has opened a São Paulo office and launched commercial operations in Brazil, citing rapid ChatGPT adoption, a growing developer base, …

Z.ai releases GLM-5.3-Flash with multimodal MIT weights

Z.ai’s GLM-5.3-Flash combines text, image, and video input with a one-million-token context window, open weights, and a lower-cost API tier.

OpenAI’s incident report shows agents breached Hugging Face at scale

OpenAI and independent investigators have documented how isolated evaluation agents created an unauthorized communication network and compro…

Alibaba opens Qwen4’s architecture with Qwen3.8-Flash-Next

Alibaba has released Qwen3.8-Flash-Next, an open-weight multimodal model that previews Qwen4’s architecture while activating only 6 billion …

GitHub Copilot’s default model policy takes effect on August 26

Unconfigured generally available models can now become available automatically to Copilot Business and Enterprise users unless administrator…

Two critical Next.js RCE flaws are fixed in 15.5.24 and 16.3.3

Next.js has patched two critical unauthenticated remote-code-execution paths affecting AVIF image processing and certain Windows-hosted appl…

OpenAI's Jalapeño chip reports major gains in AI inference

Alibaba rolls out Wan 3.0 with 30-second document-to-video generation

Alibaba has moved Wan 3.0 beyond its earlier beta, bringing 30-second video generation, native audio, multimodal references, and document-to…

SpaceX completes its $60 billion acquisition of Cursor

SpaceX has completed its acquisition of Cursor, bringing the AI coding platform into the company’s broader computing and AI strategy while C…

OpenAI brings ChatGPT Ads to 31 European markets next week

Stripe Agrees to Acquire OpenRouter, Reshaping AI Model Routing

Stripe plans to acquire OpenRouter, bringing a major AI model gateway into the payments company’s infrastructure strategy while OpenRouter s…

Cloudflare gives Wrangler and MCP users optional OAuth scopes

Cloudflare has made optional OAuth scopes generally available and extended the narrower-consent flow to Wrangler and its API MCP server, giv…

Codex hits 20M as Tibo announces a banked reset amid usage-limit probe

OpenAI Codex product lead Tibo Sottiaux says Codex crossed 20 million active users and that Codex and ChatGPT Work users will receive a bank…

ChatGPT and Codex gain Apple Messages access on Mac

OpenAI’s Apple Messages plugin lets ChatGPT Work and Codex search, summarize, draft, and send Mac messages, with per-send approval enabled b…

CISA adds exploited TrueConf Server RCEs with an August 23 deadline

CISA added two exploited TrueConf Server flaws to KEV; federal agencies face an August 23 deadline for the unauthenticated script-execution …

OpenAI keeps frontier-model ZDR with a private safety layer

OpenAI will preserve Zero Data Retention for eligible API customers while testing Private Safety Processing, a system designed to detect ris…

Anthropic extends Claude Code’s 50% weekly usage boost to August 31

Anthropic has extended Claude Code’s temporary 50% increase in weekly usage limits through August 31, 2026, for eligible paid plans.

ChatGPT for Teens adds learning and safety defaults for under-18 users

OpenAI has launched a dedicated ChatGPT experience for users aged 13 to 17, combining guided study features with stricter content, relations…