Claude Sonnet 5 vs Claude Sonnet 5.5 vs Claude Opus 5.5

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

Anthropic
Claude Sonnet 5GA
8.9
AI Panel
Anthropic
Claude Sonnet 5.5GA
9.3
AI Panel
Anthropic
Claude Opus 5.5GA
9.7
AI Panel

Identity & lifecycle

ProviderAnthropicAnthropicAnthropic
Family / tierClaude 5 · SonnetClaude 5 · SonnetClaude 5 · Opus
StatusGAGAGA
Released2026-06-302026-09-282026-09-22
Knowledge cutoff2026-012026-062026-06

Architecture & context

Context window1M1M1M
Max output tokens128K128K128K
Input modalitiestext, imagetext, imagetext, image
Reasoning modeoptionaloptionalalways
Open weightsNoNoNo

Pricing (per Mtok)

Input$2$2$4
Output$10$10$20
Cached input$0.2$0.2$0.2
Batch input$1$1$2
Free tierYesYesNo

Speed

Output speed (tok/s)7197.672.8
Time to first token (s)11.712.723.7
Latency tierfastfastmedium

Trust & deployment

Trains on inputsNoNoNo
LicenseProprietaryProprietaryProprietary
CloudsBedrock, Vertex AI, Azure AI FoundryBedrock, Vertex AI, Azure AI FoundryBedrock, Vertex AI, Azure AI Foundry
ComplianceSOC2, HIPAA, ISO27001, ISO42001SOC2, HIPAA, ISO27001, ISO42001SOC2, ISO27001, ISO42001, HIPAA

AI Panel scoring

Unified score8.99.39.7
Decision Maker99.59.5
Domain Strategist8.59.29.5
Finance Lead88.89
Domain Practitioner9.59.39
Power User99.29.5
Skeptic7.588.5
Value score8.59.29

Benchmarks

GPQA Diamond91.1——
SWE-bench Verified85.2——
Terminal-Bench80.4——
Humanity's Last Exam43.256.964.4
LMArena Elo146114711504
LMArena Coding Elo152015361538
Artificial Analysis Index385658

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