Skip to content
AI Primer

Claude Mythos Preview

A general-purpose frontier language model focused on advanced cybersecurity and autonomous coding research preview access.

Claude Mythos Preview is an April 7, 2026 Anthropic gated research-preview frontier large language model. Anthropic describes it as a general-purpose language model with especially strong cybersecurity, coding, reasoning, autonomy, and research-assistance capabilities, withheld from general availability and provided to limited defensive-security partners through Project Glasswing.

Pricing

Official site · Aug 8, 2026, 1:05 PM
Input / 1M
$10.00
Output / 1M
$50.00
Cached input / 1M
$1.00

Limited availability through Project Glasswing/trusted access. Prompt cache hits/refreshes are listed at $1/MTok; cache writes are $12.50/MTok for 5-minute cache and $20/MTok for 1-hour cache.

Anthropic's current docs identify Claude Mythos 5 as the successor/latest update to Claude Mythos Preview and direct migration from Claude Mythos Preview to Claude Mythos 5. They explicitly state Claude Mythos 5 pricing at $10 USD per million input tokens and $50 USD per million output tokens; the pricing page lists the same rates for Claude Mythos 5 with limited availability via Project Glasswing. No separate current first-party rate card for the earlier Preview name was found, so this uses Anthropic's current successor pricing rather than third-party historical claims.

View source

Model Intelligence

Context window
1,000,000 tokens
Benchmarkable
Yes
Model level
release

Recent stories

0 linked stories
No linked stories yet.
AI PrimerAI Primer

Your daily guide to AI tools, workflows, and creative inspiration.

© 2026 AI Primer. All rights reserved.