Skip to content

Claude Opus 5.5 vs Claude Sonnet 5.5

Claude Opus 5.5 is from Anthropic, released September 22, 2026. Claude Sonnet 5.5 is from Anthropic, released September 28, 2026.

  • LMArena Elo1509 vs 1477. Claude Opus 5.5 is 2.2% ahead on LMArena Elo.
  • Input per 1M$4 vs $2. Claude Sonnet 5.5 costs 50% less per million input tokens.
  • Output per 1M$20 vs $10. Claude Sonnet 5.5 costs 50% less per million output tokens.
  • Context window1M vs 1M. Both hold 1M tokens of context.

Prices are the vendors’ published API rates; Elo is LMArena’s text board. Each figure is dated on its model page.

Interactive Comparison

Loading chart...
$4 / $20 per million tokens
$2 / $10 per million tokens
Identity
DeveloperAnthropicAnthropic
ReleasedSep 2026Sep 2026
StatusGAGA
LicenceProprietaryProprietary
Self-hostableNoNo
Cost
Blended $/1M tokensinput × 0.75 + output × 0.25$8.000 / 1M tokens$4.000 / 1M tokens
Input price$4.000 / 1M tokens$2.000 / 1M tokens
Output price$20.000 / 1M tokens$10.000 / 1M tokens
Cached input$0.20 per 1M tokens$0.20 per 1M tokens
Batch discount50% ($2 / $10 per 1M tokens)50% ($1 / $5 per 1M tokens)
Capacity
Context window1M tokens1M tokens
Max output128k tokens128k tokens (up to 300k on the Message Batches API with the output-300k-2026-03-24 beta header)
Capability
Vision inYesYes
Audio inNoNo
Function callingYesYes
Structured outputYesYes
Extended reasoningYesYes - adaptive thinking, default effort high
Web searchYesYes
Code executionYesYes
Access
APIYesYes
Chat appYesYes
Cloud marketplacesAmazon Bedrock, Google Cloud, Microsoft FoundryAmazon Bedrock, Google Cloud, Microsoft Foundry
Fine-tuningNoNo
Measured quality
LMArena Elo1509 (checked Nov 2026)1477 (checked Nov 2026)
LMArena rankRank 3Rank 31
GPQA Diamond91 · max effort96 · max effort
Elo per dollarLMArena Elo ÷ blended $/1M tokens189369

Add a third model to this comparison