Trending Now - Anthropic Unveils Mythos AI to Strengthen Cybersecurity Defenses

WHO:

CNN - Anthropic has made the code of its Mythos model available to major cybersecurity and software firms, including Amazon, Apple, Cisco, Google, JPMorgan Chase, and Microsoft.

VentureBeat - Newton Cheng, Frontier Red Team Cyber Lead at Anthropic, stated, “We do not plan to make Claude Mythos Preview generally available due to its cybersecurity capabilities.”

VentureBeat - Anthropic has partnered with twelve major technology and finance companies, including Amazon Web Services, Apple, Broadcom, Cisco, CrowdStrike, Google, JPMorgan Chase, the Linux Foundation, Microsoft, Nvidia, and Palo Alto Networks, for Project Glasswing.

CNN - Logan Graham, head of the team at Anthropic responsible for AI models’ defenses, stated, “If models are going to be this good — and probably much better than this — at all cybersecurity tasks, we need to prepare pretty fast.”

CNN - Gadi Evron, founder of AI security firm Knostic, noted, “Unlike attackers, defenders don’t yet have AI capabilities accelerating them to the same degree.”

VentureBeat - Newton Cheng, Frontier Red Team Cyber Lead at Anthropic, stated, “We do not plan to make Claude Mythos Preview generally available due to its cybersecurity capabilities.”

CNN - Logan Graham, head of the team at Anthropic responsible for AI models’ defenses, stated, “We did not feel comfortable releasing this generally. We think that there’s a long way to go to have the appropriate safeguards.”

VentureBeat - Newton Cheng, Frontier Red Team Cyber Lead at Anthropic, mentioned, “However, given the rate of AI progress, it will not be long before such capabilities proliferate, potentially beyond actors who are committed to deploying them safely.”

CNN - Anthropic has briefed senior US officials “across the US government” on Mythos’ full offensive and defensive cyber capabilities.

CNN - Other firms that will have access to Mythos include chipmakers Broadcom and Nvidia, and cybersecurity vendors CrowdStrike and Palo Alto Networks.

CNN - Anthropic, a San Francisco-based AI startup, is leading the initiative.

CNN - Logan Graham, head of the team at Anthropic responsible for AI models’ defenses, commented on the release.

VentureBeat - Newton Cheng, Frontier Red Team Cyber Lead at Anthropic, discussed the model's capabilities.

VentureBeat - Elia Zaitsev, CTO of CrowdStrike, and Amy Herzog, AWS Vice President and CISO, provided insights on testing Mythos.

VentureBeat - Jim Zemlin, CEO of the Linux Foundation, highlighted the initiative's significance.

WHAT:

CNN - Mythos has autonomously found “thousands” of previously unknown software vulnerabilities, a rate far outpacing human researchers.

VentureBeat - Anthropic has committed to donating $2.5 million to Alpha-Omega and OpenSSF through the Linux Foundation, and $1.5 million to the Apache Software Foundation.

VentureBeat - The pricing for Claude Mythos Preview will be $25 per million input tokens and $125 per million output tokens after the research preview period.

VentureBeat - Anthropic has committed up to $100 million in usage credits for Project Glasswing, along with $4 million in direct donations to open-source security organizations.

VentureBeat - Mythos Preview has autonomously identified thousands of high-severity zero-day vulnerabilities, including some in every major operating system and web browser.

CNN - Anthropic has briefed senior US officials “across the US government” on Mythos’ full offensive and defensive cyber capabilities.

CNN - Mythos has autonomously found “thousands” of previously unknown software vulnerabilities, a rate far outpacing human researchers.

VentureBeat - Anthropic has committed to donating $4 million to open-source security organizations as part of Project Glasswing.

VentureBeat - Anthropic announced that it has extended access to Mythos to more than 40 additional organizations that build or maintain critical software.

VentureBeat - Mythos Preview has autonomously found a 27-year-old vulnerability in OpenBSD, a 16-year-old vulnerability in FFmpeg, and chained vulnerabilities in the Linux kernel to escalate access.

VentureBeat - Mythos Preview scored 93.9% on SWE-bench Verified, compared to 80.8% for Claude Opus 4.6, and 77.8% on SWE-bench Pro, versus 53.4% for Opus 4.6.

CNN - Anthropic will make the code of its Mythos model available to some of the world’s biggest cybersecurity and software firms to help slow the arms race ignited by AI in the hands of hackers.

CNN - Mythos has autonomously found thousands of previously unknown software vulnerabilities, a rate far outpacing human researchers.

VentureBeat - Mythos Preview scored 83.1% on the CyberGym evaluation benchmark, compared to 66.6% for Claude Opus 4.6.

VentureBeat - The model achieved 93.9% on SWE-bench Verified versus 80.8% for Opus 4.6, and 77.8% on SWE-bench Pro compared to 53.4% for Opus 4.6.

VentureBeat - Anthropic has donated $2.5 million to Alpha-Omega and OpenSSF through the Linux Foundation, and $1.5 million to the Apache Software Foundation.

Techmeme - Anthropic announced that it will make a preview of its Mythos model available to more than 40 organizations as part of Project Glasswing.

VentureBeat - Mythos Preview has autonomously identified thousands of high-severity zero-day vulnerabilities, including some in every major operating system and web browser.

VentureBeat - Mythos Preview scored 93.9% on SWE-bench Verified, compared to 80.8% for Claude Opus 4.6, and 77.8% on SWE-bench Pro, versus 53.4% for Opus 4.6.

VentureBeat - Anthropic has committed to donating $4 million to open-source security organizations as part of Project Glasswing.

VentureBeat - Anthropic has committed up to $100 million in usage credits for Project Glasswing, along with $4 million in direct donations to open-source security organizations.

VentureBeat - Mythos Preview autonomously found a 27-year-old vulnerability in OpenBSD, a 16-year-old vulnerability in FFmpeg, and chained vulnerabilities in the Linux kernel to escalate access.

CNN - Anthropic claims Mythos has found “thousands” of previously unknown software vulnerabilities at a rate far outpacing human researchers.

VentureBeat - Anthropic disclosed that its annualized revenue run rate has surpassed $30 billion as of April 5th, 2026, up from approximately $9 billion at the end of 2025.

VentureBeat - The number of business customers each spending over $1 million annually now exceeds 1,000, doubling in less than two months.

CNN - Anthropic will make the code of its Mythos model available to some of the world’s biggest cybersecurity and software firms.

CNN - Mythos, officially dubbed “Claude Mythos Preview,” is not ready for a public launch due to concerns about potential abuse by cybercriminals and spies.

VentureBeat - Mythos Preview has autonomously found a 27-year-old vulnerability in OpenBSD, a 16-year-old vulnerability in FFmpeg, and chained vulnerabilities in the Linux kernel to escalate access.

VentureBeat - Anthropic has donated $2.5 million to Alpha-Omega and OpenSSF through the Linux Foundation, and $1.5 million to the Apache Software Foundation.

VentureBeat - The pricing for Claude Mythos Preview will be $25 per million input tokens and $125 per million output tokens after the research preview period.

CNN - Anthropic announced the release of its new AI model, Mythos, for cybersecurity purposes.

CNN - Mythos has found “thousands” of previously unknown software vulnerabilities.

VentureBeat - Mythos Preview autonomously identified high-severity zero-day vulnerabilities in every major operating system and web browser.

VentureBeat - Anthropic committed up to $100 million in usage credits for Project Glasswing and $4 million in donations to open-source security organizations.

VentureBeat - The model scored 83.1% on the CyberGym evaluation benchmark.

WHERE:

CNN - The initiative involves major technology and finance companies, including Amazon, Apple, Cisco, Google, JPMorgan Chase, Microsoft, and others.

VentureBeat - The Linux Foundation is also involved in the initiative.

WHEN:

VentureBeat - Anthropic disclosed that its annualized revenue run rate has surpassed $30 billion as of April 5th, 2026, up from approximately $9 billion at the end of 2025.

CNN - The announcement of Mythos and Project Glasswing was made on April 5th, 2026.

VentureBeat - Anthropic disclosed that its annualized revenue run rate has surpassed $30 billion as of April 5th, 2026, up from approximately $9 billion at the end of 2025.

VentureBeat - The number of business customers each spending over $1 million annually now exceeds 1,000, doubling in less than two months.

CNN - The announcement of Mythos and Project Glasswing was made on April 5th, 2026.

VentureBeat - Anthropic disclosed that its annualized revenue run rate has surpassed $30 billion as of April 5th, 2026, up from approximately $9 billion at the end of 2025.

CNN - The announcement of Mythos and Project Glasswing was made on April 5th, 2026.

VentureBeat - Anthropic disclosed that its annualized revenue run rate surpassed $30 billion as of April 5th, 2026.

WHY:

CNN - Anthropic aims to give defenders a durable advantage in the coming AI-driven era of cybersecurity.

VentureBeat - Anthropic aims to give defenders a durable advantage in the coming AI-driven era of cybersecurity.

CNN - The initiative aims to strengthen cybersecurity defenses against AI-powered attacks.

VentureBeat - Anthropic considers Mythos too dangerous for public release due to its offensive capabilities.

HOW:

VentureBeat - Anthropic has built a triage pipeline specifically to manage the disclosure of vulnerabilities found by Mythos, ensuring only high-quality reports are sent to maintainers.

VentureBeat - Anthropic has built a triage pipeline to manage the disclosure of vulnerabilities found by Mythos, ensuring only high-quality reports are sent to maintainers.

CNN - Mythos will be made available to select companies for testing and evaluation to find and patch vulnerabilities.

VentureBeat - Anthropic has built a triage pipeline to manage the disclosure of vulnerabilities found by Mythos.

(UPDATED April 8, 2026 17:20 UTC)

Anthropic has made the code for its Mythos model available to major cybersecurity and software firms like Amazon, Apple, and Google, while also announcing that it will not be generally available due to its cybersecurity capabilities. The Mythos model has already identified thousands of previously unknown software vulnerabilities at a rate that far exceeds human researchers, and Anthropic plans to donate $2.5 million to Alpha-Omega and OpenSSF, along with $1.5 million to the Apache Software Foundation. As of on Sunday, Anthropic reported an annualized revenue run rate exceeding $30 billion, a significant increase from approximately $9 billion at the end of 2025.

(UPDATED April 8, 2026 15:19 UTC)

Anthropic has announced a significant partnership with twelve major technology and finance companies, including giants like Amazon Web Services, Apple, and Google, for their initiative known as Project Glasswing. The company is committing up to $100 million in usage credits and has already donated $4 million to open-source security organizations, while their Mythos Preview has autonomously identified thousands of high-severity zero-day vulnerabilities across major operating systems and web browsers. As of on Sunday, Anthropic's annualized revenue run rate has surged past $30 billion, with the number of business customers spending over $1 million annually now exceeding 1,000, effectively doubling in less than two months.

(UPDATED April 8, 2026 13:19 UTC)

on Sunday, Anthropic announced the launch of Mythos, an AI system that has autonomously identified thousands of previously unknown software vulnerabilities, significantly outpacing human researchers. Logan Graham, head of Anthropic's AI models’ defenses, emphasized the urgent need for defenders to enhance their capabilities as AI-driven cybersecurity evolves, while Gadi Evron from Knostic pointed out that defenders currently lack the same AI acceleration as attackers. In addition to briefing senior US officials on Mythos' capabilities, Anthropic has pledged $4 million to support open-source security organizations through Project Glasswing.

(UPDATED April 8, 2026 11:16 UTC)

Anthropic has announced that it has extended access to its Mythos Preview to over 40 additional organizations that develop or maintain critical software, although they do not plan to make it generally available due to its cybersecurity capabilities. The Mythos Preview has autonomously identified significant vulnerabilities, including a 27-year-old flaw in OpenBSD and a 16-year-old vulnerability in FFmpeg, while achieving impressive scores on SWE-bench Verified and Pro. Additionally, Anthropic disclosed that its annualized revenue run rate has now surpassed $30 billion as of on Sunday, a substantial increase from approximately $9 billion at the end of 2025.

(UPDATED April 8, 2026 09:17 UTC)

Anthropic has announced that it will make the code of its Mythos model available to major cybersecurity and software firms to help combat the growing threat of AI-driven hacking. The Mythos model has already demonstrated its capabilities by autonomously identifying thousands of previously unknown software vulnerabilities and achieving impressive scores on various evaluation benchmarks. In addition to this initiative, Anthropic has also donated $2.5 million to Alpha-Omega and OpenSSF through the Linux Foundation, along with $1.5 million to the Apache Software Foundation, while implementing a triage pipeline to ensure that only high-quality vulnerability reports are disclosed.

(UPDATED April 8, 2026 07:17 UTC)

Anthropic has announced that it will make a preview of its Mythos model available to over 40 organizations as part of Project Glasswing, which has already autonomously identified thousands of high-severity zero-day vulnerabilities across major operating systems and web browsers. In performance tests, Mythos Preview achieved a score of 93.9% on SWE-bench Verified, significantly outperforming Claude Opus 4.6, which scored 80.8%, and also excelled on SWE-bench Pro with 77.8% compared to Opus 4.6's 53.4%. Additionally, Anthropic has pledged to donate $4 million to open-source security organizations and has established a triage pipeline to ensure that only high-quality vulnerability reports are sent to maintainers.

(UPDATED April 8, 2026 01:20 UTC)

Anthropic has announced a commitment of up to $100 million in usage credits for Project Glasswing, along with $4 million in direct donations to open-source security organizations. Logan Graham, head of the team at Anthropic responsible for AI models’ defenses, expressed concerns about releasing certain capabilities, emphasizing the need for more safeguards, while Newton Cheng noted that the rapid progress in AI could soon lead to these capabilities being used by less responsible actors. Additionally, Anthropic's Mythos Preview has autonomously discovered several significant vulnerabilities, including a 27-year-old flaw in OpenBSD and a 16-year-old flaw in FFmpeg, and has developed a triage pipeline to ensure that only high-quality vulnerability reports are sent to maintainers.

(UPDATED April 7, 2026 23:09 UTC)

Anthropic has briefed senior U.S. officials on the full offensive and defensive cyber capabilities of its new technology, Mythos, which has reportedly identified thousands of previously unknown software vulnerabilities at a pace that far exceeds human researchers. The company also announced that its annualized revenue run rate has surpassed $30 billion as of on Sunday, with the number of business customers spending over $1 million annually now exceeding 1,000, having doubled in less than two months. In addition, Anthropic has established a triage pipeline to manage the disclosure of vulnerabilities found by Mythos, ensuring that only high-quality reports are sent to software maintainers.

(UPDATED April 7, 2026 21:11 UTC)

Anthropic has announced that it will make the code for its Mythos model available to major cybersecurity and software firms, although the model, officially named "Claude Mythos Preview," is not ready for public launch due to concerns about potential misuse. The model has already demonstrated its capabilities by autonomously discovering several vulnerabilities, including a 27-year-old flaw in OpenBSD and a 16-year-old vulnerability in FFmpeg. To support the cybersecurity community, Anthropic has donated $2.5 million to Alpha-Omega and OpenSSF, along with $1.5 million to the Apache Software Foundation, and plans to charge $25 per million input tokens and $125 per million output tokens after the research preview period.

(UPDATED April 7, 2026 19:18 UTC)

Anthropic has announced the release of its new AI model, Mythos, designed specifically for cybersecurity, which has already identified thousands of previously unknown software vulnerabilities. The initiative, involving major tech and finance companies like Amazon, Apple, and Microsoft, aims to bolster defenses against AI-powered attacks, while Anthropic has committed up to $100 million in usage credits for Project Glasswing and $4 million in donations to open-source security organizations. Mythos will be provided to select companies for testing, as Anthropic considers its offensive capabilities too dangerous for public release.