Released date

Model

Last Updated 5/22/2025

Claude Sonnet 4 (Nonthinking)

Anthropic's latest-generation workhorse model, offering a balance of performance and speed.

Released Date: 5/22/2025

Avg. Accuracy:

67.0%

Latency:

255.29s

Performance by Benchmark

Benchmarks

Accuracy

Rankings

FinanceAgent

43.5%

( 8 / 34 )

43.5%

8 / 34

CorpFin

59.9%

( 21 / 45 )

59.9%

21 / 45

CaseLaw

74.0%

( 7 / 28 )

74.0%

7 / 28

TaxEval

73.5%

( 31 / 62 )

73.5%

31 / 62

MortgageTax

74.6%

( 13 / 35 )

74.6%

13 / 35

Math500

90.3%

( 22 / 54 )

90.3%

22 / 54

AIME

38.5%

( 30 / 52 )

38.5%

30 / 52

MGSM

93.0%

( 5 / 55 )

93.0%

5 / 55

LegalBench

81.5%

( 15 / 77 )

81.5%

15 / 77

MedQA

90.3%

( 21 / 58 )

90.3%

21 / 58

GPQA

69.4%

( 21 / 54 )

69.4%

21 / 54

MMLU Pro

79.4%

( 21 / 52 )

79.4%

21 / 52

MMMU

72.4%

( 14 / 32 )

72.4%

14 / 32

LiveCodeBench

59.7%

( 25 / 54 )

59.7%

25 / 54

IOI

6.5%

( 6 / 14 )

6.5%

6 / 14

SWE-bench

65.0%

( 2 / 16 )

65.0%

2 / 16

Academic Benchmarks

Proprietary Benchmarks (contact us to get access)

Cost Analysis

Input Cost

$3.00 / M Tokens

Output Cost

$15.00 / M Tokens

Input Cost (per char)

N/A

Output Cost (per char)

$6.73 / M chars

Performance by Benchmark

Cost Analysis

Join our mailing list to receive benchmark updates on