Claude Opus 4.1 vs Claude Sonnet 4.5

Verdict

Claude Sonnet 4.5 leads 13–5 across 19 shared benchmarks.

Claude Opus 4.1 5 · Claude Sonnet 4.5 13 · 1 tied · higher isn't always better — check the caveats below.

Per-benchmark head-to-head

Reported scores — protocols may differ. How we compare

Sort

Claude Opus 4.1Claude Sonnet 4.5

2.15
AChess Puzzles
5.3
42
BCybench
60
48.3
BDeepResearch Bench
52.6
66.67
ADTBench
72
7.94
AEBR-bench
2.38
2.44
AFrontierMath-Tier-4-v2-Private
2.44
12.63
AFrontierMath-Tiers-1-3-v2-Private
23.86
43.6
AGDPval
42.5
69.7
BGPQA diamond
76.43
7.06
AHLE
9.37
43.62
ALMCA
45.6
66.81
BMETR Time Horizons
67.38
12.98
AMystery Game Puzzles
8.57
68.86
BOTIS Mock AIME 2024-2025
77.76
52
ASimpleBench
45.16
73.35
BSWE-Bench verified
71.28
38
BTerminal Bench
46.52
2.5
AVPCT
9.7
45.86
AWeirdML
47.71

Capabilities

Unknown

Unknown: Claude Sonnet 4.5 leads 3–2 across 5 shared unknown benchmarks (largest gap: EBR-bench, 7.94 vs 2.38 — single evaluator).

4 reproduced6 unverified

Reasoning

Reasoning: Claude Sonnet 4.5 leads 3–1 across 4 shared reasoning benchmarks (largest gap: SimpleBench, 52 vs 45.16 — single evaluator).

2 reproduced6 unverified

Math

Math: Claude Sonnet 4.5 leads 1–0 across 2 shared math benchmarks (largest gap: OTIS Mock AIME 2024-2025, 77.76 vs 68.86 — single evaluator).

4 reproduced

A higher score does not always mean a better model. Each number is task performance under a disclosed harness. Of the 38 measurements across 19 shared benchmarks, 38 single-evaluator or undisclosed-protocol reports; 39% are independently reproduced. Claude Sonnet 4.5 pricing is cross-check disputed — weigh cost claims with the same caution as benchmark wins. Check any benchmark's integrity grade before reading a win as decisive.

Frequently asked questions

Claude Sonnet 4.5 leads 13–5 across 19 shared benchmarks. Claude Opus 4.1 leads 5 benchmarks and Claude Sonnet 4.5 leads 13, with 1 tied. Higher isn't always better — see the integrity caveats on each benchmark.

Claude Opus 4.1 and Claude Sonnet 4.5 have 19 benchmarks in common in our data — those are the only rows where a direct, apples-to-apples comparison is drawn.

Claude Sonnet 4.5 leads reasoning 3–1. Claude Sonnet 4.5 leads unknown 3–2.

Follow the record

These models

Same tier