Hacker ApprovedDepartments

Claude Mythos

Ranked 13 of 72 language models by recommenders

Recommended for finding subtle, chained security vulnerabilities in hardened code like curl and Linux, though critics call much of it marketing hype and note restricted, expensive access.

Recommended for finding subtle, chained security vulnerabilities in hardened code like curl and Linux, though critics call much of it marketing hype and note restricted, expensive access.

511 people gave an opinion

recommend it
197
criticize it
314
approval share
39%
Each figure is 5 people. RecommendsCriticizes

Recommended in 8 of 13 months

  1. November 2025: not recommended
  2. December 2025: recommended
  3. January 2026: not recommended
  4. February 2026: not recommended
  5. March 2026: not recommended
  6. April 2026: recommended
  7. May 2026: recommended
  8. June 2026: recommended
  9. July 2026: recommended
  10. August 2026: recommended
  11. September 2026: recommended
  12. October 2026: recommended
  13. November 2026: not recommended (month in progress)

Why it doesn’t carry the badge

  • 197 people recommend itNeeds 10 or more, met.
  • 0.6 recommenders for every criticNeeds 2 to 1 or better, not met.
  • Recommended in 8 of 13 monthsNeeds 4 or more, met.

The same three rules apply to every product, and the badge can’t be bought. How the badge works

What people say about it

The reasons people give, with the number of comments that give each one. Open a reason to read what they said.

Why people recommend it

Each mark is one comment

  • Finds subtle, chained vulnerabilities

    “The Mythos reports have several examples of chaining a whole bunch of logic in different parts of the program together to exploit something very subtle.”

    Comment from 7 May 2026 in “Mozilla says 271 vulnerabilities found by Mythos and "almost no false positives"”

    “Mythos Preview is every bit as capable. So far we've found no category or complexity of vulnerability that humans can find that this model can't.”

    Comment from 21 Apr 2026 in “Mozilla: Anthropic's Mythos found 271 security vulnerabilities in Firefox 150”

    “These models are definitely a lot better than your run of the mill human developer at finding security flaws in existing systems.”

    Comment from 23 Jun 2026 in “Will It Mythos?”
  • Useful hit rate despite hallucinations

    “its hit rate on finding potential bugs and filtering out the real ones from the hallucinations is good enough to matter”

    Comment from 8 May 2026 in “Mozilla says 271 vulnerabilities found by Mythos and "almost no false positives"”

    “That they were able to use it for security scanning puts the false positive rate at a useable level, inherently.”

    Comment from 13 Apr 2026 in “Small models also found the vulnerabilities that Mythos found”

    “can greatly streamline bug hunting and exploit developments in the hands of a competent security engineer”

    Comment from 18 May 2026 in “Project Glasswing: what Mythos showed us”
  • Worth it for critical software

    “Using a very expensive model like Mythos for some little used non-critical software is overkill, but for Curl, it absolutely isn't.”

    Comment from 11 May 2026 in “Mythos Finds a Curl Vulnerability”

    “The fact that Mythos found only one vulnerability is testament to how solid curl is, not how bad Mythos is.”

    Comment from 11 May 2026 in “Mythos Finds a Curl Vulnerability”
  • Makes executives take security seriously

    “Mythos is good for cybersecurity simply because now executives can't just tell people that only superhackers can break their stuff”

    Comment from 6 May 2026 in “Mythos is the best cybersecurity news we've ever had”

    “If Mythos is what it takes to get you (your boss?) to take these issues serious, then the hype is worth it.”

    Comment from 5 Aug 2026 in “Where was Mythos when WordPress fell?”

What people warn about

Each mark is one comment

  • Overhyped marketing and scaremongering

    “Mythos looks like a good incremental model but the PR team has done a great job of associating themselves with the current trend”

    Comment from 11 May 2026 in “Mythos Finds a Curl Vulnerability”

    “Mythos simply hasn't lived up to the "we're all gunna die from the flood of vulnerabilities" hype even slightly”

    Comment from 11 May 2026 in “Mythos Finds a Curl Vulnerability”

    “Anthropic's BS marketing is something to behold and viewing it with extreme skepticism is smart.”

    Comment from 11 May 2026 in “Mythos Finds a Curl Vulnerability”
  • Restricted access, very expensive

    “Mythos is unavailable for non-privileged users and even when it will be available for more people it will be much more expensive than the alternatives.”

    Comment from 10 Apr 2026 in “US summons bank bosses over cyber risks from Anthropic's latest AI model”

    “Ever since the Mythos announcement it's been clear that we're heading towards a future where SOTA models are no longer available to the average person”

    Comment from 10 Jun 2026 in “AWS Bedrock to require sharing data with Anthropic for Mythos and future models”

    “Anthropic has been touting how their newest model is basically a cyber weapon and is so dangerous that they need to only role it out to trusted people”

    Comment from 27 Jun 2026 in “The US lifts its block on Mythos 5”
  • Not clearly better than alternatives

    “I see no evidence that this setup [Mythos] finds issues to any particular higher or more advanced degree than the other tools have done before Mythos.”

    Comment from 22 May 2026 in “Project Glasswing: An Initial Update”

    “There is little doubt that Mythos is a more powerful model, but there is no quantum leap towards Mythos”

    Comment from 13 Apr 2026 in “The Closing of the Frontier”

    “I would expect Mythos to be able to find vulnerabilities without pointing it out for him, otherwise it's no better from other agents.”

    Comment from 19 May 2026 in “Project Glasswing: what Mythos showed us”
  • Many false positives and noise

    “it is not perfect, very much similar to what pentesters would do with the added bloat of noise in reports about nonissues”

    Comment from 15 May 2026 in “Cut Off: why access to frontier AI will soon be scarce and selective”

    “80% of the security vulnerabilities that Mythos initially flagged in the Linux kernel turned out to be false positives”

    Comment from 2 Oct 2026 in “Several vulnerabilities have been discovered in the Linux kernel”

    “The trick with Mythos wasn't that it didn't hallucinate nonsense vulnerabilities, it absolutely did.”

    Comment from 11 Apr 2026 in “Small models also found the vulnerabilities that Mythos found”

Other language models people recommend

The four most recommended other language models, with the same counts for each, so you can compare like for like. The tick on each bar marks the 2 to 1 line the badge needs.

Claude Mythos compared with 4 alternatives
ProductRecommendCriticizeApprovalMonthsOwnedBadge
Claude Opus3,3312,74655%134 mo156 owners—
OpenAI GPT1,20598755%136 mo47 owners—
Claude Sonnet85458659%136 mo46 owners—
DeepSeek V4 Flash82822379%81 mo24 ownersApproved
Claude MythosThis page19731439%8——
See all 72 language models, ranked

What people said

40 of the 579 comments that mention it, newest first. Each links to the original comment.

“recognizing that bugcount alone wasn't super-relevant but also remember being impressed by an NFS bug and a kernel bug”

MixedComment from 2 Oct 2026 in “Greg Kroah-Hartman – Security in the LLM Age [video]”

“80% of the security vulnerabilities that Mythos initially flagged in the Linux kernel turned out to be false positives”

MixedComment from 2 Oct 2026 in “Several vulnerabilities have been discovered in the Linux kernel”

“Just a few months ago almost no one here believed Mythos could actually be as good at hacking as the company claimed.”

MixedComment from 28 Sep 2026 in “AI companies in race to demonstrate their model most threatening to humanity”

“If Mythos is what it takes to get you (your boss?) to take these issues serious, then the hype is worth it.”

RecommendsComment from 5 Aug 2026 in “Where was Mythos when WordPress fell?”

“Mythos is extreme hype. We are at a combo of authoritarian politicians peddling fear for power and tech bros trying to extract maximum investment returns.”

CriticizesComment from 28 Jun 2026 in “Asian AI startups launch Mythos-like models”

“Anthropic has been touting how their newest model is basically a cyber weapon and is so dangerous that they need to only role it out to trusted people”

CriticizesComment from 27 Jun 2026 in “The US lifts its block on Mythos 5”

“These models are definitely a lot better than your run of the mill human developer at finding security flaws in existing systems.”

MixedComment from 23 Jun 2026 in “Will It Mythos?”

“I think it's likely the Mythos self-restriction was a responsible action initially suggested by their AI safety team in good faith.”

MixedComment from 17 Jun 2026 in “The hacker sent by Anthropic to calm the government's nerves about AI safety”

“Ever since the Mythos announcement it's been clear that we're heading towards a future where SOTA models are no longer available to the average person”

CriticizesComment from 10 Jun 2026 in “AWS Bedrock to require sharing data with Anthropic for Mythos and future models”

“What did you expect exactly from a "10 pages epic rhyming poem about a haircut where every word starts with the letter S"?”

RecommendsComment from 10 Jun 2026 in “What it feels like to work with Mythos”

“For something like Mythos in 2026, where now the social implications of better LLMs are more understood, it's more fair to call it a marketing gimmick.”

CriticizesComment from 9 Jun 2026 in “GPT-2: Too Dangerous To Release (2019)”

“I see no evidence that this setup [Mythos] finds issues to any particular higher or more advanced degree than the other tools have done before Mythos.”

MixedComment from 22 May 2026 in “Project Glasswing: An Initial Update”

“if you do that, you get something capable of end to end chaining and reasoning about a much broader class of vulnerabilities”

RecommendsComment from 19 May 2026 in “The last six months in LLMs in five minutes”

“I would expect Mythos to be able to find vulnerabilities without pointing it out for him, otherwise it's no better from other agents.”

CriticizesComment from 19 May 2026 in “Project Glasswing: what Mythos showed us”

“can greatly streamline bug hunting and exploit developments in the hands of a competent security engineer”

MixedComment from 18 May 2026 in “Project Glasswing: what Mythos showed us”

“Weird take to claim "generally intelligent frontier" (whatever rhat means) and restrict availability based on "offensive" cyber security alone”

CriticizesComment from 15 May 2026 in “Too dangerous or just too expensive? The real reason Anthropic is hiding Mythos”

“it is not perfect, very much similar to what pentesters would do with the added bloat of noise in reports about nonissues”

MixedComment from 15 May 2026 in “Cut Off: why access to frontier AI will soon be scarce and selective”

“My personal conclusion can however not end up with anything else than that the big hype around this model so far was primarily marketing.”

CriticizesComment from 12 May 2026 in “Mythos Finds a Curl Vulnerability”

“Anthropic's BS marketing is something to behold and viewing it with extreme skepticism is smart.”

MixedComment from 11 May 2026 in “Mythos Finds a Curl Vulnerability”

“The fact that Mythos found only one vulnerability is testament to how solid curl is, not how bad Mythos is.”

MixedComment from 11 May 2026 in “Mythos Finds a Curl Vulnerability”

“Mythos looks like a good incremental model but the PR team has done a great job of associating themselves with the current trend”

MixedComment from 11 May 2026 in “Mythos Finds a Curl Vulnerability”

“Using a very expensive model like Mythos for some little used non-critical software is overkill, but for Curl, it absolutely isn't.”

MixedComment from 11 May 2026 in “Mythos Finds a Curl Vulnerability”

“Mythos simply hasn't lived up to the "we're all gunna die from the flood of vulnerabilities" hype even slightly”

MixedComment from 11 May 2026 in “Mythos Finds a Curl Vulnerability”

“It could be another marketing stunt like Mythos, which is so dangerous to release that Antrhopic must be bailed out by the government.”

CriticizesComment from 9 May 2026 in “Bun's experimental Rust rewrite hits 99.8% test compatibility on Linux x64 glibc”

“its hit rate on finding potential bugs and filtering out the real ones from the hallucinations is good enough to matter”

RecommendsComment from 8 May 2026 in “Mozilla says 271 vulnerabilities found by Mythos and "almost no false positives"”

“the most dangerous model (Mythos) which can find CVE in other projects wasn't able to find this CVE within the claude-code project itself”

MixedComment from 8 May 2026 in “Claude Code CVE-2026-39861:sandbox escape via symlink”

“The Mythos reports have several examples of chaining a whole bunch of logic in different parts of the program together to exploit something very subtle.”

RecommendsComment from 7 May 2026 in “Mozilla says 271 vulnerabilities found by Mythos and "almost no false positives"”

“Mythos did seem to be fairly clever about putting together things from different domains to find problems.”

RecommendsComment from 7 May 2026 in “Mozilla says 271 vulnerabilities found by Mythos and "almost no false positives"”

“Mythos is good for cybersecurity simply because now executives can't just tell people that only superhackers can break their stuff”

RecommendsComment from 6 May 2026 in “Mythos is the best cybersecurity news we've ever had”

“Mythos Preview is every bit as capable. So far we've found no category or complexity of vulnerability that humans can find that this model can't.”

RecommendsComment from 21 Apr 2026 in “Mozilla: Anthropic's Mythos found 271 security vulnerabilities in Firefox 150”

“They way they've published hashes of the bugs it has found so that once those bugs are fixed they can responsibly disclose them”

RecommendsComment from 20 Apr 2026 in “NSA is using Anthropic's Mythos despite blacklist”

“40 different companies that have all committed resources to patching their systems based on vulnerabilities found by Mythos”

RecommendsComment from 16 Apr 2026 in “AI cybersecurity is not proof of work”

“There is little doubt that Mythos is a more powerful model, but there is no quantum leap towards Mythos”

MixedComment from 13 Apr 2026 in “The Closing of the Frontier”

“That they were able to use it for security scanning puts the false positive rate at a useable level, inherently.”

RecommendsComment from 13 Apr 2026 in “Small models also found the vulnerabilities that Mythos found”

“The most "moral" of all moral companies, Anthropic, may not be able to stop building capabilities that look very much like weapons.”

MixedComment from 12 Apr 2026 on Hacker News

“The trick with Mythos wasn't that it didn't hallucinate nonsense vulnerabilities, it absolutely did.”

MixedComment from 11 Apr 2026 in “Small models also found the vulnerabilities that Mythos found”

“if their model can reliably write an exploit for novel bugs (starting from a crash, not a vulnerable line of code) then it's very significant”

MixedComment from 11 Apr 2026 in “Has Mythos just broken the deal that kept the internet safe?”

“With no independent reviews on the research and with many security researchers and experts accusing them of blatant scaremongering.”

CriticizesComment from 11 Apr 2026 in “Has Mythos just broken the deal that kept the internet safe?”

“Mythos is unavailable for non-privileged users and even when it will be available for more people it will be much more expensive than the alternatives.”

MixedComment from 10 Apr 2026 in “US summons bank bosses over cyber risks from Anthropic's latest AI model”

“The interesting thing about Mythos is its ability to find security vulns in software that has an uncompromised supply chain.”

RecommendsComment from 8 Apr 2026 in “Project Glasswing: Securing critical software for the AI era”

Counts come from 1,795,640 Hacker News comments, November 2025 to November 2026. How we count. Report a mistake.