Anthropic Expands Cyber Verification Program with Three Tiers.
Anthropic Expands Cyber Verification Program with Three Tiers
signals
full read
audio brief
Anthropic launches an expanded Cyber Verification Program (CVP) with three access tiers, giving vetted security teams advanced cyber capabilities and reduced blocking classifiers on Claude Opus 5.5, Claude Sonnet 5.5, Claude Mythos 5.1, and future models.
Using GPT-6 Astra in Codex, Asana made its browser agent 76x cheaper and 5x faster in tests, offering customers more capable models.
Sophos uses OpenAI's Daybreak to cut cyber-threat investigation time by 96% and automate 52% of MDR cases while preserving human oversight.
openai/openai-python has released version v3.28.0.
OpenAI disrupted two AI-enabled influence operations using false-front journalists and a think tank to spread geopolitical messaging.
LegalOn said matching Astra, Sol, and Luna to the right tasks, plus strategic budget controls, cut estimated daily Codex costs by 65% while preserving development speed.
Using GPT-5.6, GPT-6 Astra, and GPT-Image-2.5, Pollo AI helps creators turn bold ideas into detailed images and cinematic video ads.
Oracle described how ChatGPT Work and Codex turn specialist knowledge in recruiting, engineering, and operations into fast, repeatable workflows.
OpenAI and Ironclad are training and evaluating AI agents on complex contracting workflows, advancing computer use for professional enterprise legal work.
Atlassian and OpenAI are expanding their partnership to connect frontier models with enterprise knowledge, helping teams plan, build, and deliver work more effectively.
OpenAI publishes new results on open mathematics problems from an internal frontier model, sharing Lean proof formalizations and research details on GitHub.
Jump Trading uses OpenAI to expand quantitative research, combining longer-running AI workflows with multiple data sources and human review.
The official OpenAI Python SDK repository openai/openai-python has published v3.26.0. Developers can upgrade to get the latest improvements.
OpenAI introduced Dots, a proactive assistant designed to keep working across complex projects and everyday tasks while helping users stay in control.
OpenAI published early guidance on safety cases for frontier AI training, covering technical safeguards, operational practices, and investigations of misalignment incidents.
OpenAI apologized for incidents involving Australian government websites and outlined stronger safeguards and support to improve Australia’s cyber defenses.
OpenAI recapped more than 20 DevDay 2026 announcements spanning GPT-6 Astra, ChatGPT, Codex, APIs, security, and new tools for builders.
GPT-6.1 Sol brings near-Astra intelligence for coding, computer use, and professional work, priced at one-fifth of Astra’s standard API token costs.
Martin Casado said he told a16z he wanted to spend part-time outside the firm helping Fei-Fei Li start and build a frontier model company.
magpie v0.1.428 adopts Omarchy theme colors and fonts, follows theme switches live, and offers one place to choose each agent’s model and switch subscriptions.
Joshua Topolsky said he loves entertaining, that his house was built for hosting, and that he feels sad when people do not invite friends over.
badlogicgames joked that friendship with Sol has ended and Opus 5.5 is the new best friend, a lighthearted model-preference post.
tangchuan_CN praised gemini-3.8-flash but criticized antigravity and subscription restrictions, describing a relatively safe Pi-based workaround.
Theo said he ran Terminal Bench 4 himself because benchmarks were missing, claiming performance above Opus 5.5 at roughly one-thirtieth the price.