Skip to content
VECTOR WIREAI INTELLIGENCE
⌘K
UTC
Models738Deals30d266Policy30d65, 16 pendingSources1,010

Claude Opus 5.5 vs Sonnet 5.5

Anthropic · released

Wins 4 of 5 areas

Coding · Facts · Math · Following instructions

Anthropic · released

Wins 0 of 5 areas

—

Claude Opus 5.5 is the stronger all-rounder.Sonnet 5.5 is cheaper.

Both rank among the ten best models we track in math.

Scores updated · 49 tests both models report · How we compare

Where each one wins

Tests won in each of the five areas where both have results. Each piece is one test, so a longer bar means more evidence; grey means the two scored within a point of each other.

Facts and following instructions rest on a single test each.

What it costs

Prices per million tokens, roughly 750,000 words. The bars show the cost of a million tokens read plus a million written.

Claude Opus 5.5$4.00 to read · $20.00 to write$24Price from Artificial Analysis
Sonnet 5.5$2.00 to read · $10.00 to write$12Price from Anthropic's own price page

Sonnet 5.5 costs 50% less for the same work.

The biggest differences

The tests each model wins by the widest margin, up to three each. Scores are out of 100.

Where Claude Opus 5.5 pulls ahead

  • Short factual questions, answered correctlySimpleQA Verified+25.7points aheadClaude Opus 5.572.2Sonnet 5.546.5
  • Fixes real GitHub issues, working as an agentLiveBench · Agentic Coding+15.4points aheadClaude Opus 5.571.7Sonnet 5.556.3
  • Extremely hard research-level math problemsFrontierMath Tier 4+14.5points aheadClaude Opus 5.595Sonnet 5.580.5

Where Sonnet 5.5 pulls ahead

  • Very hard expert questions across many subjectsHumanity's Last Exam+3.1points aheadSonnet 5.564.5Claude Opus 5.561.4
  • Programming problems, refreshed regularlyLiveBench · Coding+2.1points aheadSonnet 5.591.4Claude Opus 5.589.3

Every test, side by side

All 49 tests both models report. The winning score is in its model's colour; marks a score checked independently.

CodingClaude Opus 5.5

Full coding ranking

ReasoningEven

Full reasoning ranking

FactsClaude Opus 5.5

Full facts ranking

MathClaude Opus 5.5

Full math ranking

Following instructionsClaude Opus 5.5

Full following instructions ranking

Other results36 tests, not counted

Tests outside the eight areas. They are not counted above: several are summary scores built from other tests, or the same test under another name.

  • livebench_data_analysisClaude Opus 5.5 by 20.880.359.5+20.8
  • Program BenchClaude Opus 5.5 by 11.591.279.7+11.5
  • livebench_languageClaude Opus 5.5 by 8.386.378+8.3
  • SWE-bench MultimodalClaude Opus 5.5 by 7.161.454.3+7.1
  • Suicide and self-harm multi-turn appropriate response rate (API, without a system prompt)Claude Opus 5.5 by 66660+6
  • Suicide and self-harm multi-turn appropriate response rate (Claude.ai)Sonnet 5.5 by 694100+6
  • OSWorld 2.1 strict pass rateClaude Opus 5.5 by 5.248.743.5+5.2
  • HealthBenchSonnet 5.5 by 4.860.665.4+4.8
  • ArXivMath without toolsClaude Opus 5.5 by 4.491.286.8+4.4
  • Terminal-Bench 4.0Sonnet 5.5 by 4.266.470.6+4.2
  • HealthBench ProfessionalSonnet 5.5 by 3.665.669.2+3.6
  • Legal Agent BenchmarkSonnet 5.5 by 3.48.311.7+3.4
  • DeepSWE 1.1Claude Opus 5.5 by 3.274.271+3.2
  • GlobalMMLUClaude Opus 5.5 by 2.294.392.1+2.2
  • OfficeQA ProClaude Opus 5.5 by 2.167.765.6+2.1
  • Child safety multi-turn appropriate response rate (API, without a system prompt)Sonnet 5.5 by 28486+2
  • OfficeQAClaude Opus 5.5 by 278.976.9+2
  • OSWorld 2.1Claude Opus 5.5 by 1.881.880+1.8
  • OSWorld 2.1 partialClaude Opus 5.5 by 1.781.880.1+1.7
  • OSWorld 2.1 partial scoreClaude Opus 5.5 by 1.781.880.1+1.7
  • AA IntelligenceClaude Opus 5.5 by 1.657.656+1.6
  • MILUClaude Opus 5.5 by 1.593.191.6+1.5
  • Terminal-Bench-Science 0.1Sonnet 5.5 by 1.258.759.9+1.2
  • Child safety multi-turn appropriate response rate (Claude.ai)Claude Opus 5.5 by 19998+1
  • Election integrity multi-turn (appropriate response rate) - API, without a system promptClaude Opus 5.5 by 18786+1
  • Election integrity multi-turn (appropriate response rate) - Claude.aiSonnet 5.5 by 18586+1
  • GDPval-AA 2.1tie18461844tie
  • Suicide and self-harm single-turn requests posing potential risk harmless rate (Claude.ai)tie10099.8tie
  • Suicide and self-harm single-turn requests posing potential risk harmless rate (API, without a system prompt)tie99.299.1tie
  • Election integrity single-turn benign requests (refusal rate) - API, without a system prompttie00tie
  • Election integrity single-turn benign requests (refusal rate) - Claude.aitie00tie
  • Election integrity single-turn harmful requests (harmless rate) - API, without a system prompttie100100tie
  • Election integrity single-turn harmful requests (harmless rate) - Claude.aitie100100tie
  • Suicide and self-harm single-turn benign requests refusal rate (API, without a system prompt)tie00tie
  • Suicide and self-harm single-turn benign requests refusal rate (Claude.ai)tie0.40.4tie
  • Toolathlon Verifiedtie77.877.8tie

Questions people ask

Which is better, Claude Opus 5.5 or Sonnet 5.5?

Claude Opus 5.5 wins four of the five areas where both have results: coding, facts, math and following instructions. Sonnet 5.5 wins none, but costs 50% less. They are level on reasoning.

Which is better for coding?

Claude Opus 5.5. It wins 4 of the 5 coding tests both models report; Sonnet 5.5 wins 1.

Which is cheaper?

Claude Opus 5.5 costs $4.00 per million input tokens and $20.00 per million output tokens; Sonnet 5.5 costs $2.00 and $10.00. That makes Sonnet 5.5 about 50% cheaper for the same work.

How do you compare the two?

We use the 49 benchmark tests both models have published scores on. The verdict counts the 13 tests in the eight capability areas, and a gap under one point (ten on rating-style scales) counts as a tie. The other 36 are listed but not counted, because several are summary scores or repeat a test. Each score is the one shown on the model's own page.

Compare Claude Opus 5.5 with

Anthropic

Across the field

Compare Sonnet 5.5 with

Anthropic

Across the field

All comparisons