RISE Humanities Data Benchmark, 0.6.1

Benchmark Results

Filter by benchmark title
 

Nursing Staff Duty Rosters

Evaluates models' ability to extract structured schedule data from visual nursing staff duty rosters, including shift types, half-day splits, alternate unit assignments, and staff metadata.

Image

Top 5 Runs

ScoreDateProviderModel
96.043 weeks agoopenaigpt-6-astra
95.841 month agoopenaigpt-6-astra
95.391 week agoanthropicclaude-opus-5-5
94.961 week agoanthropicclaude-sonnet-5-5
94.501 week agoopenaigpt-6-sol

Last 5 Runs

ScoreDateProviderModel
95.391 week agoanthropicclaude-opus-5-5
94.961 week agoanthropicclaude-sonnet-5-5
94.001 week agoanthropicclaude-fable-5-1
83.471 week agoscicoreQwen3.8-Flash-Next-FP8
1.391 week agoanthropicclaude-sonnet-5-5

Tags
  • Type(s): roster
  • Benchmark task(s):  information-extraction, document-understanding
  • Writing: printed, handwritten
  • Source creation (century): 21
  • Source Layout: table
  • Language(s): de

Personnel Cards

Evaluates models' ability to transcribe and interpret personnel index cards of Swiss federal employees (1941–1961), containing typed and handwritten entries on job title, work location, pay grade, salary, and related notes in German and French.

Image

Top 5 Runs

ScoreDateProviderModel
98.173 months agoscicoreqwen35-397b-a17b-fp8
98.172 weeks agoopenroutermeta/muse-spark-1.3
98.154 weeks agoopenroutermeta/muse-spark-1.3
98.092 weeks agoopenroutermeta/muse-spark-1.2
98.041 month agoopenroutermeta/muse-spark-1.2

Last 5 Runs

ScoreDateProviderModel
97.651 week agoanthropicclaude-sonnet-5-5
96.051 week agoscicoreQwen3.8-Flash-Next-FP8
97.241 week agoanthropicclaude-opus-5-5
97.451 week agoanthropicclaude-fable-5-1
97.461 week agoanthropicclaude-sonnet-5-5

Tags
  • Type(s): index-card
  • Benchmark task(s):  transcription, document-understanding, data-correction
  • Writing: handwritten, typed, printed
  • Source creation (century): 20
  • Source Layout: table, form
  • Language(s): de, fr