“On the darker side, it really does arm attackers to a greater degree than Mythos’s predecessors.”
Our reading, the sentence that was checked:
Claude Mythos Preview gives cyber attackers more capability than earlier Claude models did.
Assessment: Supported Evidence as of
Where and when Gary Marcus wrote it
Marcus on AI, Claude Mythos, evaluated, paragraph 3.
- Said
- Source observed
- live page, full text, public
- Evidence as of
- the cutoff for this assessment
- Assessment recorded
- assessment marcus-claude-mythos-evaluated-3c1326-a1
We recorded 3 claims from this post. See every claim from this post.
The wording above is the version this site observed on 2026-10-28. The source may have been edited after 2026-04-13. This page does not claim the wording was captured then.
What the claim commits to
- Kind
- Observation
- Deadline, as the speaker wrote it
- None stated. This site does not add one.
- Deadline state when assessed
- No horizon
- Deadline state on 2026-10-29
- No horizon (the release’s reference date)
From the speaker
- really does arm attackers
- to a greater degree
- than Mythos’s predecessors
Our interpretation
- The ordinary reading is that Mythos Preview has stronger offensive cybersecurity capabilities than earlier Claude models.
- The surrounding discussion of an unreleased model makes this a capability comparison, not a claim that criminals already have widespread access.
- I assess the current capability state described by the April 13, 2026 evaluation cited in the post.
Counts as supported if: Comparable offensive cybersecurity evaluations show a meaningful improvement over earlier Claude models in completing attacks or producing exploits.
Counts as contradicted if: Comparable evaluations show no meaningful improvement, or show that the apparent improvement disappears when testing conditions are made comparable.
Retrospective: these criteria were written after the outcome was public.
Current assessment
Was it true for the period described?
Outcome: Supported
The dated evidence matches what the claim committed to, under the criteria frozen before grading.
Scored on a plain reading. The speaker named no measure for this claim, so it was judged on the reading an informed reader would take, stated below under our interpretation. Both reviews had to find that reading fair.
- Evidence as of
- Deadline state when assessed
- No horizon
- Assessment recorded
- Policy version
- 2026-10-28.7
Rationale
AISI's common attack simulation shows a meaningful gain over earlier Claude models, including the first complete successes. This supports Marcus's capability comparison.
Limitations
- The finding concerns controlled offensive testing, not observed criminal deployment.
- The test environments were easier to attack than defended real-world networks.
Evidence (1 item)
Supports the claim: UK AI Security Institute,
the first model to solve TLO
Why it matters here: On the same 32-step attack simulation, Mythos averaged 22 completed steps, versus 16 for the next-best model, Claude Opus 4.6. Its first end-to-end completions establish stronger offensive capability in this evaluation.
- Publisher
- UK AI Security Institute
- Kind of source
- Independent measurement
- Describes events of
- Public from
Where it was said (1)
Said once in the sources this register covers.
: Marcus on AI, Claude Mythos, evaluated
On the darker side, it really does arm attackers to a greater degree than Mythos’s predecessors.
- Where
- paragraph 3
- Published
- 2026-04-13
- Source observed
- , live page
- Access
- full text, public
History
Assessment versions
- Assessment marcus-claude-mythos-evaluated-3c1326-a1: Current. Outcome: Supported. Evidence as of 2026-10-28. Public date not recorded.
Recorded events
Assessment publishedBy this site
First assessment published after both reviews and the final-text check passed.