New Finance Agent Benchmark Released

Llama 3.1 Instruct Turbo, 8B parameters with FP8 quantization.

Released Date: 7/23/2024

Avg. Accuracy:

55.1%

Latency:

2.07s

Performance by Benchmark

Benchmarks

Accuracy

Rankings

CorpFin

43.5%

( 36 / 39 )

CaseLaw

80.3%

( 28 / 62 )

ContractLaw

63.4%

( 48 / 69 )

TaxEval

39.0%

( 49 / 49 )

Math500

44.4%

( 44 / 45 )

LegalBench

52.5%

( 60 / 67 )

MedQA

62.6%

( 38 / 47 )

Academic Benchmarks
Proprietary Benchmarks (contact us to get access)

Cost Analysis

Input Cost

$0.18 / M Tokens

Output Cost

$0.18 / M Tokens

Input Cost (per char)

$0.04 / M chars

Output Cost (per char)

$0.06 / M chars

Join our mailing list to receive benchmark updates on

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