Claude Opus 4.7 vs Claude Opus 4.8

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

Anthropic
Claude Opus 4.7GA
9.0
AI Panel
Anthropic
Claude Opus 4.8GA
9.2
AI Panel

Identity & lifecycle

ProviderAnthropicAnthropic
Family / tierClaude 4 · OpusClaude 4 · Opus
StatusGAGA
Released2026-04-152026-05-27
Knowledge cutoff2026-01

Architecture & context

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

Pricing (per Mtok)

Input$5$5
Output$25$25
Cached input$0.5$0.5
Batch input$2.5$2.5
Free tierNoNo

Speed

Output speed (tok/s)54.660
Time to first token (s)23.850.5
Latency tierslowmedium

Trust & deployment

Trains on inputsNoNo
LicenseProprietaryProprietary
CloudsBedrock, Vertex AI, Azure AI FoundryBedrock, Vertex AI, Azure AI Foundry
ComplianceSOC2, HIPAA, GDPR, ISO27001SOC2, HIPAA, GDPR, ISO27001

AI Panel scoring

Unified score99.2
Decision Maker99
Domain Strategist99
Finance Lead88.5
Domain Practitioner9.59.5
Power User8.59
Skeptic7.58
Value score7.58

Benchmarks

GPQA Diamond94.293.6
SWE-bench Verified87.688.6
MMMU91.5
Terminal-Bench69.474.6
Humanity's Last Exam54.7
LMArena Elo1503
LMArena Coding Elo1554
Artificial Analysis Index57

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