claude-mythos-5.1

Vendor anthropic Tier frontier Price per million tokens, in / out $10 / $50 Released 2026-09-01 Context 1,000,000 Max output 128,000 Vendor page

Last updated 2026-10-26 How this is made

Meant to excel at

No sourced claims yet.

Watch out for

No sourced claims yet.

Benchmarks

Agentic coding percentile 0.88

Benchmark	Score	Percentile	Source	Date
terminal-bench 4.0	60.9	0.88	Vendor claim, relayed handyai.substack.com	2026-09-01

Percentile is the share of tracked models this model beats on that benchmark, averaged across the group. Only benchmarks at least four tracked models share are scored.

Practitioner reports

+/-

Zscaler announced a limited preview where Mythos 5.1 powers an 'Exploit Paths' feature that links evidence into attack paths with confidence levels; the post gives no measured performance figures.

Task: operations-support | 2026-10-05
Vendor claim, relayed zscaler.com
+/-

Launch-day HN commenters said Mythos 5.1 is effectively the same model as Fable 5.1 with different safeguards, so the Terminal-Bench gap (60.9 vs 55.8) is mostly classifier interventions rather than capability.

Task: operations-support | Compared with claude-fable-5.1 | 2026-09-08
Practitioner report eesel.ai