AI Rankings

Anthropic

Claude Mythos Preview

Anthropic's invite-only cybersecurity research preview offered through Project Glasswing, deprecated on the Claude API since June 2026.

Visit official site (opens in a new tab)Compare with another model

Our editorial assessment, based on the public sources listed below. Last reviewed .

Overall rank
#33
Tier
A
Score
82.4
Main category
Coding

Strengths

  • Per Anthropic, its coding and reasoning results were well beyond earlier Claude models at release
  • Used by Glasswing partners for defensive work such as finding and fixing vulnerabilities
  • Per Anthropic, it found thousands of zero-day vulnerabilities, including in major operating systems and browsers

Considerations

  • Not generally available: access is invite-only through Project Glasswing, with no Claude consumer plan
  • Deprecated on the Claude API since June 2026; Anthropic now offers Mythos 5 and 5.1 under limited access

Price

Invite-only via Project Glasswing; $25 input / $125 output per 1M tokens after credits

Shown for reference — price is not part of the score.

Score breakdown

Performance
90.0
55% of score
Reputation
84.0
35% of score
Recency
35.0
10% of score

The overall score combines these sub-scores using the weights shown. How we score →

Verification status

Each field carries its own re-verification date. Amber means older than 90 days — treat those figures as stale until re-checked.

Pricing
Benchmark
Reputation
source ↗
Feature list

Checked against the sources listed below.

Sources

Show 7 more sources

Similar models