Hugging Face's logo

Datasets:
TIGER-Lab
/
ClawBench

Tasks:
Text Generation
Question Answering
Other
Modalities:
Tabular
Text
Formats:
parquet
Languages:
English
Size:
< 1K
ArXiv:
Tags:
arxiv:2604.08523
benchmark
leaderboard
agent-benchmark
llm-benchmark
web-agents
Libraries:
Datasets
pandas
Polars
License:
Dataset card Data Studio Files
xet
 Community
3
ClawBench / leaderboard

Commit History

  1. leaderboard: add avg_cost_per_task_usd column for V2 hermes rows

    AgPerry
    Verified signature
    9055bda

  2. fix: restore deepseek-v4-flash:free Total=129 per canonical table (concurrent edit set it to 130)

    AgPerry
    Verified signature
    1c1bb0d

  3. fix: reorder V2 hermes rows to match website (owl@5, glm-air@6, dsflash@7, minimax@8); dsflash total 129->130

    AgPerry
    Verified signature
    a81baf6

  4. feat: add reward_rate_strict column for 8 V2 hermes rows (lenient/strict alignment)

    AgPerry
    Verified signature
    145660e

  5. Update V2 hermes leaderboard rows with ds-v4-pro lenient judge (2026-05-20)

    AgPerry
    Verified signature
    1bc9856

  1. V2: Hermes-only — drop glm-5.1 openclaw row (alt harnesses tracked separately)

    AgPerry
    Verified signature
    dfd621f

  1. Refresh leaderboard: 7 V2 + 6 V1 rows with reward_rate (judge-graded)

    AgPerry
    Verified signature
    387d311

  2. Refresh results.csv: 7 V2 + 6 V1 rows with reward_rate (judge-graded)

    AgPerry
    Verified signature
    238ecea

  1. Duplicate from NAIL-Group/ClawBench

    AgPerry
    c6bec71