Claude Opus 5 vs Claude Sonnet 5.5

Best cell per row highlighted. Null means undisclosed — never counted as zero.

Anthropic
Claude Opus 5GA
9.5
AI Panel
Anthropic
Claude Sonnet 5.5GA
9.3
AI Panel

Identity & lifecycle

ProviderAnthropicAnthropic
Family / tierClaude 5 · OpusClaude 5 · Sonnet
StatusGAGA
Released2026-07-242026-09-28
Knowledge cutoff2026-052026-06

Architecture & context

Context window1M1M
Max output tokens128K128K
Input modalitiestext, imagetext, image
Reasoning modeoptionaloptional
Open weightsNoNo

Pricing (per Mtok)

Input$5$2
Output$25$10
Cached input$0.5$0.2
Batch input$2.5$1
Free tierNoYes

Speed

Output speed (tok/s)—97.6
Time to first token (s)—12.7
Latency tiermediumfast

Trust & deployment

Trains on inputsNoNo
LicenseProprietaryProprietary
CloudsBedrock, Vertex AI, Azure AI FoundryBedrock, Vertex AI, Azure AI Foundry
ComplianceSOC2, ISO27001, ISO42001, HIPAA, CSA-STARSOC2, HIPAA, ISO27001, ISO42001

AI Panel scoring

Unified score9.59.3
Decision Maker9.59.5
Domain Strategist99.2
Finance Lead8.58.8
Domain Practitioner9.59.3
Power User99.2
Skeptic88
Value score8.59.2

Benchmarks

SWE-bench Verified96—
Humanity's Last Exam56.356.9
LMArena Elo14931471
LMArena Coding Elo15331536
Artificial Analysis Index—56

Scores link to their sources. Missing cells mean the vendor hasn't disclosed a result — honesty over padding.