Benchmark Almanac

Weights

We track 4 benchmarks that Weights made or helped make.

0 in use 4 with public code 1 with a leaderboard here

Benchmarks

Name Kind Released Best score Status
Nejumi LLM Leaderboard 3 Models complete Japanese language and alignment evaluations. The organizers group results by use case and include safety evaluations. Reading in many languages Jul 2024 0.83 Not in cards
Nejumi LLM Leaderboard Models complete Japanese JGLUE tasks without demonstrations. The leaderboard compares model scores on the same Japanese evaluation setup. Reading in many languages Jul 2023 Not in cards
Nejumi LLM Leaderboard 2 Models complete Japanese evaluations using llm-jp-eval datasets. The repository publishes the evaluation configurations and data preparation instructions. Reading in many languages Dec 2023 Not in cards
Nejumi LLM Leaderboard 4 Models solve Japanese language, reasoning, coding, and alignment tasks. The organizers weight individual categories when calculating the overall score. Reading in many languages Aug 2025 Not in cards