Skip to content
VECTOR WIREAI INTELLIGENCE
⌘K
UTC
Models728Deals30d254Policy30d66, 13 pendingSources985

Claude Mythos 5 vs Claude Mythos Preview

16 SHARED BENCHMARKS

Across 16 shared benchmarks, Claude Mythos 5 scores higher on 10 and Claude Mythos Preview on 6. The widest gap is RiemannBench, where Claude Mythos 5 scores 55 against 43.

ANTHROPICVSANTHROPIC16 SHARED10–6 HEAD-TO-HEADNEWEST SCORE ADDED

At a glance

Claude Mythos 5Claude Mythos Preview
MakerAnthropicAnthropic
Released—
Price per 1M tokens input / output$10.00 / $50.00—
Cost of 1M in + 1M out$60.00—
Head-to-head of 16 shared benchmarks10 wins6 wins
Scores tracked independently verified48 0 ◆25 0 ◆

Claude Mythos 5's release date per the vendor's own announcement. Prices: Anthropic's own price page for Claude Mythos 5. ◆ = independently verified score.

Where each leadsby capability area · benchmark wins

AreaClaude Mythos 5WINSClaude Mythos Preview
Reasoning01Claude Mythos Preview leads 1 of 1 · widest: Humanity's Last Exam 57.8 vs 64.7
Coding21Claude Mythos 5 leads 2 of 3
Agentic10Claude Mythos 5 leads 1 of 1 · widest: OSWorld-Verified 85 vs 79.6
Multimodal01Claude Mythos Preview leads 1 of 1 · widest: CharXiv (RQ) 88.9 vs 93.2

Biggest gaps

Claude Mythos 5 pulls furthest ahead on

  1. OSWorld-Verified85 vs 79.6

Claude Mythos Preview pulls furthest ahead on

  1. Humanity's Last Exam64.7 vs 57.8
  2. CharXiv (RQ)93.2 vs 88.9

Every shared benchmark16 · grouped by area

Multimodal 1

BenchmarkClaude Mythos 5MARGINClaude Mythos Preview
CharXiv (RQ)88.993.2

Each score is the one the model's own page shows — the most authoritative tracked result for that benchmark, on the benchmark's standard methodology (source-attributed). ↓ marks lower-is-better metrics; ◆ an independently verified score. Open either model for its full surface, provenance and pricing.