Claude Mythos 5 — benchmark results

Anthropic Claude Mythos 5 model for high-end reasoning, coding, agentic, and professional tasks. Provider: Anthropic. Released 2026-06-09. Access: API.

Unified ELO 2250 ± 31, rank #3 of 1776 rated models, from 57 benchmark results.

Strongest benchmark results

BenchmarkScoreMetricPercentile
100Q-Hard Net Score42Net score (self-reported)100
AA-Omniscience Net Score53Net score (self-reported)100
ArxivMath (Fable/Mythos)78.5Score (%)100
BenchLM83.9Overall Score100
BioMysteryBench Human Difficult (Fable/Mythos)46.1Score (%)100
BioMysteryBench Human Solvable (Fable/Mythos)83.9Score (%)100
BioMysteryBench Human-Difficult46.1Accuracy (self-reported)100
BioMysteryBench Human-Solvable83.9Accuracy (self-reported)100
BrowseComp (Fable/Mythos Single-Agent)88Score (%)100
CharXiv Reasoning (Fable/Mythos No Tools)88.9Score (%)100
CharXiv Reasoning (Fable/Mythos Tools)93.5Score (%)100
CharXiv-R93.5Score (self-reported)100

Interactive version: theaggregate.ai/model?slug=claude-mythos-5 · How the rankings work · Data refreshed daily, snapshot 2026-07-22.