Anthropic Launches Cyber Verification Program with Claude Opus 5.5

Key Takeaways

• <ul><li>Anthropic's Cyber Verification Program offers three tiers — Defense, Red Team, and Specialized Access — with the last allowing the fewest restrictions for testing critical infrastructure like power grids.</li></ul>
• <ul><li>Claude Opus 5.5 completed 34 of 50 attack scenarios under Red Team Access, matching its unrestricted 67.6 percent success rate in CyScenarioBench testing.</li></ul>
• <ul><li>Glasswing partners verified at least 129,000 vulnerabilities from April to July 2026, with over 33,000 rated critical or high severity; Anthropic expects the true total to be at least five times higher.</li></ul>
• <ul><li>The Cyber Verification Program operates on Claude Platform, Google Cloud Vertex AI, Microsoft Foundry, and Amazon Bedrock.</li></ul>
• <ul><li>Amazon blocked Meta's Muse AI agent from accessing its website over data privacy and security concerns, joining a trend of sites adding permission barriers for AI agents.</li></ul>
• <ul><li>Jim Cramer named Meta and Microsoft as his top AI stocks, citing Zuckerberg's Muse strategy and Nadella's Azure and Copilot success.</li></ul>
• <ul><li>C3.ai reported declining billings, low gross margins, and only one quarter of positive free cash flow, signaling ongoing cash burn risk.</li></ul>
• <ul><li>RSA Agent ID helps regulated industries register AI agents and verify accountability for every important decision the AI makes.</li></ul>
• <ul><li>OpenAI's ChatGPT for Teens features drove 2.7 million additional learning messages, with a new College Planner tool in development for high school students.</li></ul>
• <ul><li>55 percent of California college students say AI influenced their major choice, while 36 percent of adults expect positive job effects and 37 percent expect negative ones.</li></ul>

Anthropic Expands Access to Powerful AI Models for Security Teams

Anthropic expanded its Cyber Verification Program to give security professionals access to advanced models like Claude Opus 5.5, Claude Sonnet 5.5, and Claude Mythos 5.1. The program has three tiers covering defensive research, penetration testing, and critical infrastructure evaluation. Standard public models keep strict safeguards, but approved defenders get fewer restrictions. The goal is to balance cybersecurity testing needs with oversight to prevent misuse.

Anthropic Widens AI Cyber Access for Vetted Security Teams

Anthropic expanded its Cyber Verification Program with three tiers based on cybersecurity work type. Defense Access covers security operations and incident response. Red Team Access adds authorized penetration testing. Specialized Access has the fewest restrictions for testing critical systems like power grids and telecom networks. Anthropic tested the tiers using CyScenarioBench, where Claude Opus 5.5 completed 34 of 50 attack scenarios under Red Team Access, matching its 67.6 percent success rate without safeguards.

Anthropic Opens Mythos-Class AI to More Defenders

Anthropic merged Project Glasswing and its Cyber Verification Program into one structure with Defense, Red Team, and Specialized access tiers. Glasswing partners verified at least 129,000 vulnerabilities between April and July 2026, and Anthropic found 5,500 more through open-source scanning. Over 33,000 were rated critical or high severity. The Red Team tier completed 34 of 50 attack scenarios without blocks, matching the unrestricted model rate. The program runs on Claude Platform, Google Cloud Vertex AI, Microsoft Foundry, and Amazon Bedrock.

Anthropic Launches 3-Tier Cyber Verification Program

Anthropic introduced three access tiers: Defense Access, Red Team Access, and Specialized Access. Defense Access covers security operations, malware reverse engineering, and vulnerability validation for security teams and researchers. Red Team Access permits penetration testing only on authorized systems. Specialized Access is for a limited set of organizations testing safety-critical systems like power grids and interbank infrastructure, vetted with the US government. Glasswing partners found over 33,000 critical or high severity vulnerabilities, and Anthropic expects the true impact to be at least five times higher.

Las Vegas Radiologists Turn to AI Amid Doctor Shortages

Desert Radiology in Las Vegas is using two AI tools to address rising demand for medical imaging as the valley faces doctor shortages. The adoption of artificial intelligence helps radiologists manage increased workloads. The initiative comes as the area experiences a shortage of medical professionals. The tools are being used to support diagnostic work in radiology.

Las Vegas radiologists turn to AI amid doctor shortages

Desert Radiology in Las Vegas is using two AI tools to meet growing demand for medical imaging. The area is facing a shortage of doctors. Desert Radiology is among the first in the valley to use AI this way. The tools help manage rising imaging needs and doctor shortages.

Midwest Wheel develops AI agents to solve problems

Midwest Wheel is building AI agents that can fix problems on their own. Agentic AI governance helps the company decide which tasks software agents can handle and which need human oversight. Moving to the cloud freed the lean IT team and allowed faster tool development. The company wants AI to not just find problems but solve them, while keeping humans in charge of financial decisions.

Aston Martin uses back-end AI while avoiding customer-facing agents

Aston Martin is using AI behind the scenes to improve operations like configure-price-quote systems. The carmaker evaluates each AI idea for profit and efficiency but keeps agents off its customer website. It assigns people in the business to own each AI agent's outcomes. The company also built a governance framework to guide AI use while preserving its personalized customer experience.

C3.ai billings decline raises concerns for shareholders

C3.ai reported year over year declines in billings and low gross margins compared to software peers. The company faces pressure from higher infrastructure costs and weaker profitability. It has recorded only one quarter of positive free cash flow. Its overall free cash flow margin remains negative, pointing to ongoing cash burn risk despite heavy growth spending.

Amazon blocks Meta's Muse AI agent over data concerns

Amazon has blocked Meta's Muse AI agent from accessing its website due to data privacy and security concerns. Muse, launched in 2022, uses machine learning and natural language processing to respond to user queries. The block highlights a growing trend of websites tightening controls on AI agents. Companies like Amazon and Google are increasingly adding permission barriers for AI agent access.

Nebius Stock Rises While COO Sells Half His Shares

Nebius Group shares rose 7.4% on Tuesday as investors reacted positively to the company's growth in AI infrastructure. Chief Operating Officer Ophir Nave sold 500,000 shares worth about $118 million, which was half of his stake. This sale happened under a pre-set trading plan while other AI stocks like Nvidia and Marvell also saw gains. Retail investor sentiment for Nebius improved from bearish to neutral during this trading session.

Jim Cramer Picks Meta and Microsoft as Top AI Stocks

Jim Cramer chose Meta Platforms and Microsoft as his best AI investment opportunities on his show Mad Money. He trusts Meta CEO Mark Zuckerberg to find new ways to use their AI product called Muse. Cramer also likes Microsoft because of CEO Satya Nadella and the success of their Azure cloud services and Copilot tool. He believes these established companies can grow even when interest rates are high.

New Tools Help Teens Learn and Plan College with AI

OpenAI released new features for ChatGPT for Teens to help young people learn and prepare for college. Teens using the tool sent 2.7 million more learning messages and used study modes to understand math and science concepts. The company is adding a College Planner feature to help students track application deadlines and manage financial aid steps. These tools aim to support students in grades 10 through 12 without replacing the need for human guidance.

RSA Agent ID Helps Regulated Industries Secure AI Actions

RSA Agent ID is a new tool designed to manage identities for AI agents in strict industries like finance and government. It allows organizations to register AI agents and prove who is behind every important decision the AI makes. This system helps companies meet board-level accountability requirements and follow safety regulations. By closing the identity gap, the tool reduces the risk of security breaches caused by unverified AI actions.

California Students Say AI Changed Their Major Choices

A new report shows that 55% of California college students say AI influenced the major they chose to study. The survey found that 36% of adults believe AI will have a positive effect on their jobs while 37% expect a negative effect. More than half of these students use AI regularly for schoolwork, with 20% using it every day. These findings came out after researchers at Anthropic made controversial comments about the future risks of artificial intelligence.

Sources

• Anthropic Opens Its Most Powerful AI Models to More Security Teams
• Anthropic widens access to AI cyber capabilities for vetted security teams
• Anthropic Opens Mythos-Class AI to More Defenders
• Anthropic Introduces 3-Tier Cyber Verification Program for AI Access
• How Las Vegas radiologists are using AI as valley sees doctor shortages
• How Las Vegas radiologists are using AI as valley sees doctor shortages
• Midwest Wheel builds toward AI agents that fix problems
• Aston Martin puts back-end AI to work
• What C3.ai (AI)'s Billings Decline Means For Shareholders
• AI Agents Face Permission Barriers as Websites, OS Tighten Controls
• Nebius Stock Rallies With AI Infrastructure Peers Even As COO Cuts Stake: Retail Mood Improves
• Jim Cramer picks Meta and Microsoft as top AI stocks
• Helping teens learn, plan, and shape the future of AI
• Stop Hoping for Secure AI and Start Building It: RSA Agent ID Closes the Agentic Identity Gap for Highly Regulated Industries
• AI is impacting what California students chose to study, report says