88Z AI, 8 April — Anthropic has released the preview version of Claude Mythos, which outperforms Opus 4.6. Due to high security risks, access is restricted to giants like Amazon, Apple, and Google, as well as over 40 infrastructure companies via Project Glasswing for defensive security hardening.
At a glance
- Claude Mythos outperforms Opus 4.6 and can detect thousands of high-risk vulnerabilities
- Access is restricted to specific enterprises via Project Glasswing due to high security risks
The update
Anthropic has officially launched the preview version of its most powerful model, Claude Mythos, which significantly outperforms Opus 4.6. Tests indicate that Claude Mythos possesses strong code analysis capabilities, able to identify thousands of high-risk vulnerabilities, with potential attack skills surpassing those of most hackers. Due to the extremely high security risks associated with the model, Anthropic has decided not to make it available to the public for now. Currently, the model is only accessible to tech giants such as Amazon, Apple, and Google, as well as over 40 critical infrastructure companies, through a special initiative called Project Glasswing. These partners will use the model for defensive security hardening to enhance system safety.
Sources & editorial note
This article is an AI-assisted synthesis by 88Z AI based on the public sources below. Product claims and availability are attributed to their publishers. It does not represent independent testing or first-hand reporting.
- 量子位Source published
Source dates above refer to the original announcements; the date under the headline is this article’s publication on 88Z AI.