New Finance Agent Benchmark Released

Released date

Model

Last Updated 9/18/2025

mistralai/magistral-medium-2509

Magistral Medium 1.2 (09/2025)

Reasoning model with vision support released in September 2025.

Released Date: 9/18/2025

Avg. Accuracy:

54.3%

Latency:

278.93s

Performance by Benchmark

Benchmarks

Accuracy

Rankings

18.2%

( 28 / 42 )

18.2%

28 / 42

52.3%

( 42 / 55 )

52.3%

42 / 55

52.4%

( 37 / 38 )

52.4%

37 / 38

66.3%

( 53 / 72 )

66.3%

53 / 72

66.8%

( 33 / 44 )

66.8%

33 / 44

84.5%

( 13 / 62 )

84.5%

13 / 62

74.6%

( 61 / 65 )

74.6%

61 / 65

53.7%

( 75 / 87 )

53.7%

75 / 87

89.5%

( 30 / 68 )

89.5%

30 / 68

38.6%

( 58 / 64 )

38.6%

58 / 64

68.6%

( 49 / 62 )

68.6%

49 / 62

64.6%

( 29 / 41 )

64.6%

29 / 41

74.9%

( 12 / 61 )

74.9%

12 / 61

0.7%

( 18 / 19 )

0.7%

18 / 19

28.7%

( 14 / 23 )

28.7%

14 / 23

34.8%

( 17 / 23 )

34.8%

17 / 23

Academic Benchmarks

Proprietary Benchmarks (contact us to get access)

Cost Analysis

Input Cost

$2.00 / M Tokens

Output Cost

$5.00 / M Tokens

Join our mailing list to receive benchmark updates on

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