NSA Uses Anthropic's AI Despite Pentagon Blacklisting Company
Intra-Party Split Detected
Trump administration officials split between Pentagon hardliners who view Anthropic as untrustworthy and other agencies/White House officials who want access to the company's advanced AI capabilities
Left says
- •The government's cybersecurity needs are outweighing political disputes, as agencies recognize Anthropic's tools are essential for protecting critical infrastructure like the electric grid and financial system from cyberattacks
- •The Pentagon's blacklisting stems from Anthropic's principled refusal to allow its AI to be used for mass domestic surveillance and autonomous weapons development, showing the company prioritizes ethical guardrails
- •High-level White House meetings signal the administration recognizes that depriving the U.S. of cutting-edge AI capabilities would be 'grossly irresponsible' and 'a gift to China'
- •Intelligence agencies and civilian departments are already successfully using Anthropic's tools for legitimate national security purposes without the problematic military applications the Pentagon demanded
Right says
- •Anthropic is exploiting its powerful Mythos model as leverage to circumvent legitimate Pentagon concerns about supply chain security and the company's reliability as a defense contractor
- •The company's refusal to provide 'full, unrestricted access' for 'all lawful purposes' demonstrates it cannot be trusted to support military needs when national security is at stake
- •Anthropic is using 'fear tactics' about Mythos's cyber capabilities to find sympathetic government officials while simultaneously fighting the Pentagon's authority in court
- •The security breach allowing unauthorized access to Mythos raises serious questions about whether Anthropic can effectively safeguard its most dangerous technologies from malicious actors
Common Take
High Consensus- Anthropic's Mythos model possesses advanced cybersecurity capabilities that could be used both defensively to harden systems and offensively to exploit vulnerabilities
- The U.S. government has legitimate national security interests in accessing cutting-edge AI technology to maintain competitive advantage over adversaries like China
- Multiple government agencies beyond the Pentagon, including Treasury and intelligence services, have expressed interest in using Anthropic's AI tools
- The dispute centers on specific contractual terms about usage restrictions rather than fundamental disagreements about AI's importance to national security
The Arguments
Right argues
Anthropic's refusal to provide 'full, unrestricted access' for 'all lawful purposes' demonstrates fundamental unreliability as a defense contractor, especially when national security demands could require immediate access to AI capabilities without corporate gatekeeping.
Left counters
Anthropic's principled stance against mass domestic surveillance and autonomous weapons development shows responsible AI governance, and the company has explicitly stated it supports all other lawful military uses including cybersecurity defense.
Left argues
The government's cybersecurity needs are so urgent that agencies like NSA are already using Mythos despite Pentagon objections, proving that Anthropic's tools are essential for protecting critical infrastructure from sophisticated cyber threats.
Right counters
This unauthorized use while simultaneously fighting the Pentagon in court demonstrates Anthropic's willingness to exploit government divisions and use fear tactics about cyber capabilities to circumvent legitimate supply chain security processes.
Left argues
High-level White House meetings with Treasury Secretary Bessent and Chief of Staff Wiles signal that depriving the U.S. of cutting-edge AI capabilities would be 'grossly irresponsible' and 'a gift to China' in the global AI competition.
Right counters
Anthropic is strategically leveraging Mythos's dangerous capabilities as a bargaining chip to pressure sympathetic officials while undermining Pentagon authority, creating dangerous precedents for corporate influence over national security decisions.
Right argues
The security breach allowing unauthorized access to Mythos through a third-party vendor environment raises serious questions about whether Anthropic can effectively safeguard its most dangerous technologies from malicious actors.
Left counters
The breach occurred in a contractor environment, not Anthropic's core systems, and the company's restricted release model to only 40 trusted organizations actually demonstrates responsible security practices compared to broader public releases.
Left argues
Intelligence agencies and civilian departments like Energy and Treasury have legitimate needs to use Mythos for defensive cybersecurity purposes that don't involve the problematic military applications the Pentagon demanded.
Right counters
Anthropic cannot be trusted to make these distinctions reliably when national emergencies arise, and the company's current court challenge shows it prioritizes corporate control over government authority in national security matters.
Challenge Questions
These questions target genuine internal contradictions — meant to provoke honest reflection.
Right asks Left
“If Anthropic's ethical guardrails are truly principled rather than strategic, why is the company simultaneously fighting Pentagon restrictions in court while actively lobbying other government agencies for access, and how does this selective cooperation serve national security rather than corporate interests?”
Left asks Right
“If the Pentagon's supply chain risk designation is truly about legitimate security concerns rather than bureaucratic overreach, why are other intelligence and civilian agencies successfully using Anthropic's tools for national security purposes without the security breaches or reliability issues the Pentagon claims to fear?”
Outlier Report
Left Fringe
Progressive tech activists and AI safety advocates like Timnit Gebru or members of groups like Algorithmic Justice League who view any military AI cooperation as inherently problematic represent roughly 15% of the left coalition.
Right Fringe
Hardline national security hawks like Tom Cotton or defense industry advocates who would demand complete corporate submission to military needs without any ethical guardrails represent about 20% of the right coalition.
Noise Assessment
Moderate noise level - most discourse reflects genuine disagreement about balancing corporate ethics with national security rather than performative positioning.
Sources (9)
<p>The National Security Agency is using <a href="https://www.axios.com/2026/04/16/anthropic-claude-opus-model-mythos" target="_blank">Anthropic's</a> most powerful model yet, <a href="https://www.axios.com/2026/04/07/anthropic-mythos-preview-cybersecurity-risks" target="_blank">Mythos Preview</a>, despite top officials at the Department of Defense — which oversees the NSA — insisting the company is a "supply chain risk," two sources tell Axios. </p><p><strong>Why it matters: </strong>The government's cybersecurity needs appear to be outweighing the Pentagon's feud with Anthropic</p><hr /><ul><li>The department moved in February to cut off Anthropic and force its vendors to follow suit. That case is ongoing.</li><li>The military is now broadening its use of Anthropic's tools while simultaneously arguing in court that using those tools threatens U.S. national security.</li></ul><p><strong>Breaking it down: </strong>Two sources said the NSA was using Mythos, while one said the model was also being used more widely within the department.</p><ul><li>It's unclear how the NSA is currently using Mythos, but other organizations with access to the model are using it predominantly to scan their own environments for exploitable security vulnerabilities.</li><li>Anthropic restricted access to Mythos to around 40 organizations, contending that its offensive cyber capabilities were too dangerous to allow for a wider release.</li><li>Anthropic only <a href="https://www.anthropic.com/glasswing" target="_blank">announced 12</a> of those organizations. One source said the NSA was among the unnamed agencies with access.</li><li>The NSA's counterparts in the <a href="https://www.gov.uk/government/publications/ai-cyber-threats-open-letter-to-business-leaders/ai-cyber-threats-open-letter-to-business-leaders-html" target="_blank">U.K.</a> have <a href="https://www.yahoo.com/news/articles/mi5-called-protect-britain-breakthrough-060000124.html" target="_blank">said</a> they have access to the model through the country's AI Security Institute.</li></ul><p><strong>Driving the news: </strong>Anthropic CEO Dario Amodei <a href="https://www.axios.com/2026/04/17/anthropic-white-house-wiles-bessent-amodei" target="_blank">met</a><strong> </strong>White House chief of staff Susie Wiles and Treasury Secretary Scott Bessent on Friday to discuss the use of Mythos within government and Anthropic's wider plans and security practices. </p><ul><li>Sources said next steps after the meeting were expected to focus on how departments other than the Pentagon engage with the model. Both sides described the meeting as productive.</li><li>Anthropic and the Pentagon declined to comment. The NSA and Office of the Director of National Intelligence did not respond to requests for comment.</li></ul><p><strong>Zoom out: </strong>The breakdown between the Pentagon and Anthropic came during tense contract renegotiations earlier this year.</p><ul><li>The department demanded Anthropic make its Claude model available for "all lawful purposes," while the company insisted on walling off mass domestic surveillance and the development of autonomous weapons.</li><li>Some Defense officials still insist Anthropic's posture proved it can't be trusted to be there when the military needs it. Anthropic denies that.</li><li>Others in the administration just want this fight to go away so they can utilize the cutting edge tools Anthropic is building. </li></ul>
<p>Treasury Secretary Scott Bessent joined a <a href="https://www.axios.com/2026/04/17/anthropic-trump-administration-mythos" target="_blank">meeting</a> on Friday between White House chief of staff Susie Wiles and Anthropic CEO Dario Amodei, sources briefed on the meeting told Axios.</p><p><strong>Why it matters: </strong><a href="https://www.axios.com/2026/04/16/anthropic-claude-opus-model-mythos" target="_blank">Anthropic</a> is building tools that could have enormous implications for the federal government. But that same government is currently fighting Anthropic in court after the Pentagon declared it a "supply chain risk." The meeting points to a potential thaw.</p><hr /><p><strong>"This is a big problem. </strong>Everyone's complaining. There's all this drama. So this got elevated to Susie to hear Dario out, determine what is bullsh-t and start to plot a way forward," a Trump adviser told Axios.</p><ul><li>Going into the meeting, both sides sought to wall off the Pentagon fight from the way the rest of the government engaged with Anthropic, and next steps are expected to be about how other departments engage with Anthropic's new Mythos Preview model, according to another source familiar with the negotiations.</li><li>In the meeting, Wiles circumspect about the Pentagon aspect, saying "it's in court." But she made clear that the government needs a relationship with Anthropic and she wants an open line of communication, one source said.</li><li>The discussion covered how Anthropic is safeguarding its code and how the company makes decisions around things like when and how to release new models.</li></ul><p><strong>Driving the news: </strong>Due to its frightening cyber capabilities,<strong> </strong>Anthropic released Mythos only to a select group of companies and organizations.</p><ul><li>Treasury and other government agencies have expressed <a href="https://www.axios.com/2026/04/16/white-house-anthropic-ai-mythos-government-national-security" target="_blank">interest in joining that list</a>. Two sources told Axios prior to the White House meeting that a deal along those lines could be struck soon.</li><li>One concern is that Mythos and other cutting-edge AI tools could allow nefarious hackers to breach the U.S. financial system. Alternatively, companies and government agencies could utilize Mythos to harden their cyber defenses before bad actors get access.</li><li>Bessent joined the meeting because "he wants to make sure everyone is on the same page," a source familiar with his thinking said. "He understands this is a private company, but there is a role for government to play here."</li></ul><p><strong>What they're saying: </strong>An Anthropic spokesperson described the meeting afterwards as "a productive discussion on how Anthropic and the U.S. government can work together on key shared priorities such as cybersecurity, America's lead in the AI race, and AI safety."</p><ul><li>The White House said afterward that the "introductory meeting" had been "productive and constructive," adding: "We discussed opportunities for collaboration, as well as shared approaches and protocols to address the challenges associated with scaling this technology. The conversation also explored the balance between advancing innovation and ensuring safety."</li></ul><p><strong>Split screen:</strong> Anthropic is still fighting the "supply chain risk" designation, which the Pentagon applied after the company refused to make its software available to the military for "all lawful uses."</p><ul><li>Amodei insisted on restricting its use for mass surveillance or to develop autonomous weapons. During weeks of bitter and ultimately failed negotiations, some in the Pentagon grew to loathe Anthropic and Amodei.</li><li>President Trump declared at the time of the Pentagon designation that all government agencies must cease their use of Anthropic, but that pronouncement is on hold as the court battle continues.</li><li>The company still has plenty of detractors inside the administration. "They're using this Mythos cyber weapon to find friendly ears in the government," one U.S. official said. "They're succeeding."</li></ul><p><strong>The bottom line: </strong>Another U.S. official said that "every agency except [the Department of] War wants to" use Anthropic. </p>
<p>Anthropic CEO Dario Amodei is scheduled to walk into the West Wing on Friday for a meeting with White House chief of staff Susie Wiles — a breakthrough in his effort to resolve the company's bitter <a href="https://www.axios.com/technology/automation-and-ai" target="_blank">AI fight</a> with the Pentagon.</p><p><strong>Why it matters:</strong> The Trump administration recognizes the power of Anthropic's new Claude model, <a href="https://www.anthropic.com/glasswing" target="_blank">Mythos</a>, and its highly sophisticated — and potentially dangerous — ability to <a href="https://www.axios.com/2026/04/08/mythos-system-card" target="_blank">breach cybersecurity defenses</a>.</p><hr /><ul><li>"It would be grossly irresponsible for the U.S. government to deprive itself of the technological leaps that the new model presents," a source close to negotiations told us. "It would be a gift to China."</li></ul><p><strong>Reminder: </strong>Anthropic is <a href="https://www.axios.com/2026/04/02/trump-administration-appeals-anthropic-pentagon" target="_blank">suing the Pentagon</a> for <a href="https://x.com/SecWar/status/2027507717469049070" target="_blank">blacklisting the company</a> after Amodei <a href="https://www.anthropic.com/news/where-stand-department-war" target="_blank">refused to allow</a> his AI to be used without restrictions.</p><ul><li>Some parts of the U.S. <a href="https://www.intelligence.gov/" target="_blank">intelligence community</a>, plus the Cybersecurity and Infrastructure Security Agency (CISA, part of Homeland Security), are testing Mythos. Treasury and others want it.</li></ul><p><strong>Behind the scenes: </strong>After Anthropic took the administration to court, negotiations with the Pentagon chilled. But Anthropic has hired key Trumpworld consultants — so expect a deal. Friday's meeting is designed to pave the way.</p><p><strong>Flashback:</strong> This is the second time Amodei has held a high-stakes meeting with a top Trump official this year.</p><ul><li>In late February, Defense Secretary Pete Hegseth gave Amodei until the end of the week to accept the Pentagon's terms, or else. Anthropic didn't.</li><li>Since then, the Pentagon and Anthropic have been locked in a legal and political feud. Some in the administration think the fight is growing counterproductive.</li></ul><p>💡 <em>If you're a CEO or on a CEO's team: <a href="https://www.axios.com/newsletters/c-suite" target="_blank">Apply here</a> for free admission to Jim's new Axios C-Suite weekly newsletter. </em></p><ul><li><em><a href="https://www.axios.com/2026/04/16/white-house-anthropic-ai-mythos-government-national-security" target="_blank">Go deeper</a>: Trump officials negotiating access to Anthropic's Mythos despite blacklist.</em></li></ul>
<p>The White House and <a href="https://www.axios.com/2026/04/16/anthropic-claude-power-user-complaints" target="_blank">Anthropic</a> are in active discussions about deploying the AI firm's powerful new model, <a href="https://www.axios.com/2026/04/07/anthropic-mythos-preview-cybersecurity-risks" target="_blank">Mythos Preview</a>, within the federal government despite ongoing efforts to blacklist the company as a supply chain risk, sources familiar with the discussions tell Axios. </p><p><strong>Why it matters: </strong>Anthropic is in a bitter feud with <a href="https://www.axios.com/2026/03/26/anthropic-pentagon-ai-deal" target="_blank">the Pentagon</a>, but even U.S. officials who dislike the company concede that it's building tools that could aid U.S. national security — or harm it, if they fall into the wrong hands.</p><hr /><p><strong>Driving the news: </strong>Anthropic is only rolling out Mythos to <a href="https://www.axios.com/2026/04/07/anthropic-mythos-preview-cybersecurity-risks" target="_blank">a select group</a> of companies and organizations, rather than the general public, so they can assess its frightening cyber capabilities and harden their defenses.</p><ul><li>Some government agencies want to join that club, and the White House and Anthropic are discussing the terms under which that might be possible.</li><li>In response to queries from agencies as to whether they are able to use Mythos, the Office of Management and Budget sent out an email, first reported by Bloomberg, saying they were looking into the matter.</li><li>Two sources told Axios there are ongoing discussions with Anthropic and agencies may get access to Mythos in the coming weeks.</li></ul><p><strong>Friction point: </strong>That's despite the fact that Anthropic is in <a href="https://www.axios.com/2026/04/08/anthropic-loses-bid-to-block-pentagon-blacklisting" target="_blank">ongoing litigation with the Pentagon</a>, which declared the company a "<a href="https://www.axios.com/2026/03/09/anthropic-sues-pentagon-supply-chain-risk-label" target="_blank">supply chain risk</a>" and ordered companies that work with the military to remove its software from those workflows.</p><ul><li>Anthropic is currently barred from Pentagon contracts but can do business with the rest of the government while the case continues.</li><li>"There's progress with the White House. There's not progress with [the Department of] War," a administration official said.</li><li>A second administration official said the government "has a responsibility to evaluate every model to see where the frontier of tech is," but accused Anthropic of using "fear tactics" by issuing warnings of how Mythos could supercharge hacking.</li><li>"They're using this Mythos cyber weapon to find friendly ears in the government," the official said. "They're succeeding."</li></ul><p><strong>What they're saying: </strong>"All the intel agencies use Anthropic. Every agency except War wants to. That's because Anthropic doesn't want to kill people and War's position is 'don't tell us what the f*** to do.' But if you're the Department of Energy, you don't give a f*** about that. You're worried about the Chinese attacking the energy grid. So you want Anthropic," the first administration official said.</p><ul><li>Anthropic's official position is that it will not allow its models to be used for mass surveillance or to develop fully autonomous weapons.</li><li>The Pentagon says that's unduly restrictive because those definitions are nebulous and it needs assurances it can use AI systems for "all lawful purposes."</li><li>Most of those concerns don't apply to non-military work.</li><li>Anthropic and the Pentagon both declined to comment.</li></ul><p><strong>Zoom in:</strong> Civilian agencies like the Departments of Energy and Treasury are responsible for safeguarding critical sectors like the electric grid and financial system.</p><ul><li>Accessing Mythos would help them determine where companies and local governments may be vulnerable to cyberattacks and how to help them prepare.</li></ul><p><strong>Between the lines: </strong>Key officials in the Trump administration see Anthropic and its leaders as woke doomsters, and some relished slapping on the "supply chain risk" designation.</p><ul><li>But some of those same officials, and many others, also see Anthropic's tools as best-in-class when it comes to AI for national security purposes. </li><li>One Defense official told Axios at the height of the Pentagon-Anthropic feud that the only reason the talks were ongoing is "these guys are that good."</li></ul><p><strong>Go deeper: </strong><a href="https://www.axios.com/2026/03/26/anthropic-pentagon-ai-deal" target="_blank">How Anthropic's Pentagon deal could get revived</a></p><p><em>Sam Sabin contributed reporting.</em></p>
<p><a href="https://www.axios.com/2026/04/13/anthropic-revenue-growth-ai" target="_blank">Anthropic</a> on Thursday released Claude Opus 4.7, a meaningful upgrade to its flagship AI model with better coding, sharper vision and a new ability to double-check its own work.</p><p><strong>Why it matters: </strong>Anthropic publicly conceded that the new Opus model does not match the performance of <a href="https://www.axios.com/2026/04/07/anthropic-mythos-preview-cybersecurity-risks" target="_blank">Mythos</a>, a highly advanced system that the company hasn't released to the public due to safety concerns.</p><hr /><ul><li>In a chart accompanying its <a href="https://x.com/claudeai/status/2044785261393977612?s=20" target="_blank">announcement</a>, Anthropic showed that Opus 4.7 beats Opus 4.6, ChatGPT 5.4, Google Gemini 3.1 Pro in a number of key benchmarks.</li><li>But Opus 4.7 still falls short of its Mythos Preview model, which has only been released to a handpicked group of tech and cybersecurity companies.</li></ul><p><strong>What they're saying:</strong> "Opus 4.7 is a notable improvement on Opus 4.6 in advanced software engineering, with particular gains on the most difficult tasks," Anthropic said in a <a href="https://www.anthropic.com/news/claude-opus-4-7" target="_blank">blog post</a>. </p><ul><li>"Users report being able to hand off their hardest coding work—the kind that previously needed close supervision—to Opus 4.7 with confidence."</li><li>"The model also has substantially better vision: it can see images in greater resolution," it said. "It's more tasteful and creative when completing professional tasks, producing higher-quality interfaces, slides, and docs."</li></ul><p><strong>The big picture:</strong> The release arrives amid <a href="https://www.axios.com/2026/04/16/anthropic-claude-power-user-complaints" target="_blank">weeks of user complaints</a> that Opus 4.6 had quietly gotten worse.</p><ul><li>"Claude has regressed to the point it cannot be trusted to perform complex engineering," an AMD senior director wrote in a <a href="https://x.com/firstadopter/status/2043086033454031281?s=20" target="_blank">widely shared</a> post on GitHub.</li><li>Speculation centered on whether Claude has been deliberately scaled back — what users are calling "<a href="https://x.com/ShanuMathew93/status/2042645545970213202" target="_blank">nerfed</a>" — either to control costs or to redirect scarce compute toward Mythos and other frontier efforts.</li><li>Anthropic denied that any changes it made were to redirect computing resources to other projects.</li></ul><p><strong>Zoom in</strong>: In addition to the new model, Anthropic is shifting its default level of reasoning in Claude Code and offering a new option. </p><ul><li>"Opus 4.7 introduces a new xhigh ("extra high") effort level between high and max, giving users finer control over the tradeoff between reasoning and latency on hard problems," Anthropic said."When testing Opus 4.7 for coding and agentic use cases, we recommend starting with high or xhigh effort."</li><li>It's also testing a new system called "task budgets" that give developers more control over how Claude does its reasoning on longer tasks.</li></ul><p><strong>Between the lines</strong>: Anthropic said it will use the new release to test guardrails designed to prevent its model being used for cybersecurity attacks.</p><ul><li>"What we learn from the real-world deployment of these safeguards will help us work towards our eventual goal of a broad release of Mythos-class models," the company said.</li></ul>
<p>Anthropic users across online forums are raising the same complaint: Claude suddenly feels ... bad.</p><p><strong>Why it matters:</strong> The backlash lands just as Anthropic is <a href="https://www.axios.com/2026/04/07/anthropic-mythos-preview-cybersecurity-risks" target="_blank">testing</a> a more powerful model, Mythos — raising questions about whether cutting-edge AI is becoming less accessible even as it gets more capable.</p><hr /><p><strong>Driving the news:</strong> Over the past few weeks, users on X, GitHub and Reddit have been swapping anecdotes, benchmarks and prompts in an effort to pinpoint what changed and why.</p><ul><li>"Claude has regressed to the point it cannot be trusted to perform complex engineering," an AMD senior director wrote in a <a href="https://x.com/firstadopter/status/2043086033454031281?s=20" target="_blank">widely shared</a> post on GitHub.</li><li><a href="https://x.com/iruletheworldmo/status/2043332977136975994?s=46" target="_blank">Others</a> have posted <a href="https://x.com/bridgemindai/status/2043321284113670594?s=46" target="_blank">side-by-side</a> outputs and benchmarks they say show Claude generating answers that are <a href="https://x.com/Hesamation/status/2042979500103815306" target="_blank">less accurate </a>or <a href="https://x.com/clairevo/status/2041287600154288400" target="_blank">nuanced</a>. </li><li>Much of the speculation centers on whether Claude has been deliberately scaled back — what users are calling "<a href="https://x.com/ShanuMathew93/status/2042645545970213202" target="_blank">nerfed</a>" — either to control costs or to redirect scarce compute toward Mythos and other frontier efforts.</li></ul><p><strong>The other side:</strong> Anthropic says it adjusted the default level of reasoning in Claude Code, but denies the changes were tied to compute constraints or Mythos.</p><ul><li>When asked about the online complaints, Anthropic pointed Axios to a post on X from Boris Cherny, head of Claude Code, <a href="https://x.com/bcherny/status/2029970236460691885?s=46" target="_blank">from March 6</a>.<strong> </strong></li><li>"You can change it anytime in the /model selector if you prefer low effort (faster) or high effort (more intelligence). The setting is sticky and will persist for your next session," Cherny said.</li></ul><p><strong>Between the lines:</strong> Analyst Patrick Moorhead decided to <a href="https://x.com/PatrickMoorhead/status/2044074719888982319?s=20" target="_blank">ask Claude</a> to weigh in.</p><ul><li>"Anthropic made real configuration changes that objectively reduced default thinking depth across all surfaces including claude.ai, but the most extreme 'secret nerfing' narrative overstates what happened," Claude said as part of its lengthy response.</li></ul><p><strong>Another theory</strong> is that users aren't seeing decline so much as acclimating to what previously felt magical.</p><ul><li>Over time, expectations rise and flaws become more noticeable — a phenomenon <a href="https://x.com/daniel_mac8/status/2043676629617045811?s=20" target="_blank">known as habituation</a>. </li></ul><p><strong>Yes, but:</strong> Even if the change is explainable, the perception problem is real — especially for power users relying on consistent performance for coding and research workflows.</p><p><strong>The big picture:</strong> The fight over Claude's "intelligence" points to a broader shift: access to top-tier AI is fragmenting.</p><ul><li>Advanced capabilities are increasingly gated behind higher-cost tiers, API usage or experimental programs.</li><li>Anthropic is also reportedly close to <a href="https://www.theinformation.com/briefings/exclusive-anthropic-preps-opus-4-7-model-ai-design-tool" target="_blank">upgrading</a> its high-end Opus model to version 4.7.</li></ul><p><strong>The increasing stratification</strong> could lead to a division between those who can afford to pay top dollar for the best models and those who can't.</p><ul><li>Anthropic recently <a href="https://www.theinformation.com/articles/anthropic-changes-pricing-bill-firms-based-ai-use-amid-compute-crunch?ref=implicator.ai" target="_blank">moved</a> large enterprise customers to a fully usage-based (token) pricing model, tying intelligence more directly to spend.</li><li>It's also reinforcing a <a href="https://www.axios.com/2026/04/13/ai-elite-vs-ai-skeptic-doomer" target="_blank">widening divide</a> between power users and dabblers around AI's capabilities.</li></ul><p><strong>What we're watching:</strong> Whether "default" AI experiences continue to get worse even as frontier systems get dramatically stronger.</p>
<p>AI startup Anthropic is investigating reports that a group of users gained unauthorized access to its Claude Mythos model, a powerful AI system that was deliberately released to only a small circle of trusted companies due to its advanced cybersecurity capabilities.</p> <p>The post <a href="https://www.breitbart.com/tech/2026/04/22/ai-security-breach-anthropic-investigates-unauthorized-access-to-powerful-claude-mythos-model/" rel="nofollow">AI Security Breach: Anthropic Investigates Unauthorized Access to Powerful ‘Claude Mythos’ Model</a> appeared first on <a href="https://www.breitbart.com" rel="nofollow">Breitbart</a>.</p>
The Pentagon previously labeled Anthropic a supply-chain risk.
Anthropic’s new Mythos model is keeping the company’s foot in the White House’s door despite the Trump administration blacklisting the firm’s products from military and government work earlier this year.  Mythos, Anthropic’s most advanced model to date, has drawn interest from various parts of the federal government, giving the artificial intelligence firm a chance to…