New Finance Agent Benchmark Released

Released date

Model

Last Updated 4/5/2025

together/meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8

Llama 4 Maverick

Llama 4 Maverick 17B 128E Instruct FP8

Released Date: 4/5/2025

Avg. Accuracy:

63.3%

Latency:

7.90s

Performance by Benchmark

Benchmarks

Accuracy

Rankings

3.6%

( 20 / 22 )

3.6%

20 / 22

57.6%

( 16 / 35 )

57.6%

16 / 35

76.1%

( 33 / 55 )

76.1%

33 / 55

63.3%

( 44 / 62 )

63.3%

44 / 62

69.3%

( 24 / 42 )

69.3%

24 / 42

72.7%

( 11 / 23 )

72.7%

11 / 23

85.2%

( 14 / 38 )

85.2%

14 / 38

25.2%

( 17 / 34 )

25.2%

17 / 34

92.5%

( 5 / 36 )

92.5%

5 / 36

77.2%

( 27 / 60 )

77.2%

27 / 60

43.3%

( 40 / 40 )

43.3%

40 / 40

67.7%

( 10 / 35 )

67.7%

10 / 35

79.4%

( 11 / 35 )

79.4%

11 / 35

72.6%

( 7 / 22 )

72.6%

7 / 22

Academic Benchmarks

Proprietary Benchmarks (contact us to get access)

Cost Analysis

Input Cost

$0.27 / M Tokens

Output Cost

$0.85 / M Tokens

Input Cost (per char)

$0.06 / M chars

Output Cost (per char)

$0.20 / M chars

Join our mailing list to receive benchmark updates on

Stay up to date as new benchmarks and models are released.