Noometry

Model comparison

Claude Opus 4.6 vs Claude Sonnet 4.5

Claude Opus 4.6 is the stronger model overall, scoring 58.3 to 44.0 on the Noometry Index. Claude Sonnet 4.5 costs 1.7× less per token, which makes it the better buy when Claude Opus 4.6's lead doesn't matter for your workload.

Last verified . 58 shared benchmarks.

Claude Opus 4.6 Anthropic

58.3

Rank #18 Confirmed

Claude Sonnet 4.5 Anthropic

44.0

Rank #76 Confirmed

Summary

  • They share 58 benchmarks with published results for both. Claude Opus 4.6 scores higher in 10 categories and Claude Sonnet 4.5 in 0 categories; 10 gaps are clear of the uncertainty.
  • The widest gap is in reasoning, where Claude Opus 4.6 leads 57.7 to 26.9.
  • The biggest single-benchmark swing is ARC-AGI-2: 69.2% for Claude Opus 4.6 and 13.6% for Claude Sonnet 4.5.
  • Claude Sonnet 4.5 is cheaper at $3 / $15 per million input/output tokens, against $5 / $25 for Claude Opus 4.6.

Side by side

Claude Opus 4.6Claude Sonnet 4.5
ProviderAnthropicAnthropic
Noometry Index58.344.0
Released2026-02-042025-09-29
WeightsProprietaryProprietary
Context window1M1M
Max output128K64K
Input $ / M tokens$5$3
Output $ / M tokens$25$15
Results tracked6873

Sponsored placements are available on pages like this one. Advertise on Noometry

Category by category

Coding Claude Opus 4.6 leads

Claude Opus 4.6: 57.7 (#15), Claude Sonnet 4.5: 47.8 (#54)

BenchmarkClaude Opus 4.6Claude Sonnet 4.5
SWE-bench Verified78.7%71.3%
SWE-bench Verified (bash only)75.6%71.4%
LMArena WebDev15461393
SWE-bench Multilingual72%67%
GSO41.2%14.7%
WeirdML78%47.7%
LMArena Coding15351489
ALE-Bench996.5796.15
AlgoTune1.471.52
FrontierCode26.6%—
SciCode—44.7%

Agentic & Tool Use Claude Opus 4.6 leads

Claude Opus 4.6: 51.0 (#4), Claude Sonnet 4.5: 38.1 (#30)

BenchmarkClaude Opus 4.6Claude Sonnet 4.5
Terminal-Bench79.8%46.5%
Remote Labor Index4.2%2.1%
τ²-bench Banking27.3%25.3%
Cybench93%60%
DeepResearch Bench55.3%52.6%
LMArena Search12531159
METR Time Horizons78.9%67.4%
Vending-Bench 28,0183,839
APEX-Agents46.3%—
Berkeley Function Calling Leaderboard—73.2%
GDPval—42.5%
τ²-bench Airline—72%
τ²-bench Retail—72.4%
τ²-bench Telecom—84.9%
OSWorld—62.9%
GBAEval44.1%—

Reasoning Claude Opus 4.6 leads

Claude Opus 4.6: 57.7 (#21), Claude Sonnet 4.5: 26.9 (#113)

BenchmarkClaude Opus 4.6Claude Sonnet 4.5
ARC-AGI-269.2%13.6%
SimpleBench67.6%54.3%
Kagi LLM Benchmark83.6%57.9%
NYT Connections (extended)92.1%37.3%
ARC-AGI-194%63.7%
Chess Puzzles17%12%
EnigmaEval7.6%6%
EBR-Bench12.7%2.4%
LMArena Hard Prompts15271462
Mystery Game Puzzles25%17%
DTBench91.2%83.2%
LMCA55.8%38.8%
Epoch Capabilities Index155.36146.84
ForecastBench6061.9
CritPt—1.1%
Thematic Generalization80.6%—

Math Claude Opus 4.6 leads

Claude Opus 4.6: 62.7 (#25), Claude Sonnet 4.5: 31.9 (#190)

BenchmarkClaude Opus 4.6Claude Sonnet 4.5
FrontierMath (Tiers 1-3)66%23.9%
FrontierMath Tier 426.8%2.4%
OTIS Mock AIME 2024-202594.4%77.8%
ProofBench50%19%
LMArena Math15171448
FrontierMath (Feb 2025 set)40.7%15.2%
FrontierMath Tier 4 (v1)22.9%4.2%
MathArena Final-Answer Competitions78.5%—
Omni-MATH—55.3%
MATH Level 5—97.7%

Knowledge Claude Opus 4.6 leads

Claude Opus 4.6: 61.8 (#23), Claude Sonnet 4.5: 48.4 (#70)

BenchmarkClaude Opus 4.6Claude Sonnet 4.5
GPQA Diamond90.5%82.3%
Humanity's Last Exam34.4%13.7%
SimpleQA Verified47%30.7%
Vectara Hallucination Rate12.2%12%
LMArena Expert15461483
MMLU-Pro—86.9%
GPQA (HELM)—68.6%

Multimodal Claude Opus 4.6 leads

Claude Opus 4.6: 37.0 (#67), Claude Sonnet 4.5: 34.1 (#84)

BenchmarkClaude Opus 4.6Claude Sonnet 4.5
LMArena Document15071450
LMArena Vision1316—
VPCT—39.8%
Furniture Assembly28.3%—

Multilingual Claude Opus 4.6 leads

Claude Opus 4.6: 58.2 (#4), Claude Sonnet 4.5: 53.6 (#65)

BenchmarkClaude Opus 4.6Claude Sonnet 4.5
LMArena Non-English14901425
LMArena Chinese15511459
LMArena French15131459
LMArena German15041427
LMArena Japanese14841390
LMArena Korean14651403
LMArena Russian14971436
LMArena Spanish15121457

Instruction Following Claude Opus 4.6 leads

Claude Opus 4.6: 79.7 (#3), Claude Sonnet 4.5: 75.2 (#74)

BenchmarkClaude Opus 4.6Claude Sonnet 4.5
LMArena Instruction Following15231459
IFEval—85%

Long Context Claude Opus 4.6 leads

Claude Opus 4.6: 48.1 (#13), Claude Sonnet 4.5: 45.2 (#45)

BenchmarkClaude Opus 4.6Claude Sonnet 4.5
LMArena Longer Query15201476
CL-bench20.7%—
CL-bench Life17%—

Writing & Preference Claude Opus 4.6 leads

Claude Opus 4.6: 73.7 (#9), Claude Sonnet 4.5: 66.7 (#33)

BenchmarkClaude Opus 4.6Claude Sonnet 4.5
LMArena Text15041439
LMArena Creative Writing15061442
EQ-Bench Creative Writing18091678
LMArena Multi-Turn15141465
WildBench—85.4%
EQ-Bench 41223—

Frequently asked questions

Is Claude Opus 4.6 better than Claude Sonnet 4.5?

Claude Opus 4.6 is the stronger model overall, scoring 58.3 to 44.0 on the Noometry Index. Claude Sonnet 4.5 costs 1.7× less per token, which makes it the better buy when Claude Opus 4.6's lead doesn't matter for your workload.

Which is cheaper, Claude Opus 4.6 or Claude Sonnet 4.5?

Claude Sonnet 4.5 is cheaper. It lists at $3 per million input tokens and $15 per million output tokens; Claude Opus 4.6 lists at $5 and $25.

Is Claude Opus 4.6 or Claude Sonnet 4.5 better for coding?

Claude Opus 4.6 scores higher on coding benchmarks: 57.7 versus 47.8 in the Noometry coding category.

Which has the bigger context window?

Both accept 1M tokens.

How many benchmarks do Claude Opus 4.6 and Claude Sonnet 4.5 share?

58 benchmarks have published results for both models. Claude Opus 4.6 has 68 scored results on Noometry and Claude Sonnet 4.5 has 73.

Related comparisons

Go deeper