VECTOR WIREAI INTELLIGENCE
⌘K
UTC
Refresh Models —Deals —Regulatory —Sources —

Claude Mythos 5 vs Claude Opus 5.5

14 SHARED BENCHMARKS

Across 14 shared benchmarks, Claude Mythos 5 scores higher on 2 and Claude Opus 5.5 on 12. The widest gap is Terminal-Bench-Science 0.1, where Claude Opus 5.5 scores 58.7 against 24.7. Claude Opus 5.5 is the cheaper of the two on tracked API pricing ($4.00 against $10.00 per million input tokens).

ANTHROPICVSANTHROPIC14 SHARED2–12 HEAD-TO-HEADNEWEST SCORE ADDED

At a glance

Claude Mythos 5Claude Opus 5.5
MakerAnthropicAnthropic
Released
Price per 1M tokens input / output$10.00 / $50.00$4.00 / $20.00
Cost of 1M in + 1M out$60.00$24.00 2.5× less
Head-to-head of 14 shared benchmarks2 wins12 wins
Scores tracked independently verified51 0 ◆37 1 ◆

Release dates per the vendor's own announcement. Prices: Anthropic's own price page for Claude Mythos 5; Artificial Analysis for Claude Opus 5.5. ◆ = independently verified score.

Where each leadsby capability area · benchmark wins

AreaClaude Mythos 5WINSClaude Opus 5.5
Reasoning01Claude Opus 5.5 leads 1 of 1 · widest: Humanity's Last Exam 57.8 vs 67.7
Coding02Claude Opus 5.5 leads 2 of 2 · widest: SWE-bench Pro 80 vs 89.9

Biggest gaps

Claude Mythos 5 pulls furthest ahead on

No ratified-area lead of 3 points or more.

Claude Opus 5.5 pulls furthest ahead on

  1. Humanity's Last Exam67.7 vs 57.8
  2. SWE-bench Pro89.9 vs 80
  3. SWE-bench Multilingual93.9 vs 86.6

Every shared benchmark14 · grouped by area

Reasoning 1

BenchmarkClaude Mythos 5MARGINClaude Opus 5.5
Humanity's Last Exam57.867.7

Other shared benchmarks 11

BenchmarkClaude Mythos 5MARGINClaude Opus 5.5
AECI164.5169.4
HealthBench Professional63.365.6
OfficeQA7978.9
OfficeQA Pro67.167.7
OSWorld 2.0 (partial)72.981.8
OSWorld 2.0 (strict)36.148.7
Program Bench86.391.2
SWE-bench Multimodal54.161.4
Terminal-Bench 4.04266.4
Terminal-Bench-Science 0.124.758.7
Toolathlon Verified79.377.8

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. Quoted rates are the price-setter row we currently track for each model — its direct or vendor-official listing where one exists (anthropic-official, direct), otherwise the lowest tracked offer.