{"generated_at":"2026-05-12","links":{"dataset":"https://huggingface.co/datasets/TIGER-Lab/ClawBench","github":"https://github.com/reacher-z/ClawBench","leaderboard_space":"https://huggingface.co/spaces/TIGER-Lab/ClawBench","paper_arxiv":"https://arxiv.org/abs/2604.08523","results_csv":"https://huggingface.co/datasets/TIGER-Lab/ClawBench/resolve/main/leaderboard/results.csv","scoring_md":"https://github.com/reacher-z/ClawBench/blob/main/eval/scoring.md","website":"https://claw-bench.com/"},"rows_v1":[{"corpus":"v1","harness":"hermes","int_rate":0.6143790849673203,"intercepted":94,"matched":94,"model":"claude-opus-4-6","n":153,"partial":false,"reward":0.6143790849673203,"wall_h":null},{"corpus":"v1","harness":"hermes","int_rate":0.5686274509803921,"intercepted":87,"matched":87,"model":"claude-sonnet-4-6","n":153,"partial":false,"reward":0.5686274509803921,"wall_h":null},{"corpus":"v1","harness":"hermes","int_rate":0.3006535947712418,"intercepted":46,"matched":46,"model":"claude-haiku-4-5-20251001","n":153,"partial":false,"reward":0.3006535947712418,"wall_h":null},{"corpus":"v1","harness":"hermes","int_rate":0.2549019607843137,"intercepted":39,"matched":39,"model":"gpt-5.4-2026-03-05","n":153,"partial":false,"reward":0.2549019607843137,"wall_h":null},{"corpus":"v1","harness":"hermes","int_rate":0.24836601307189543,"intercepted":38,"matched":38,"model":"gpt-5.4-mini-2026-03-17","n":153,"partial":false,"reward":0.24836601307189543,"wall_h":null},{"corpus":"v1","harness":"hermes","int_rate":0.17647058823529413,"intercepted":27,"matched":27,"model":"kimi-k2.5","n":153,"partial":false,"reward":0.17647058823529413,"wall_h":null}],"rows_v2":[{"corpus":"v2","corpus_intercepted":0.3153846153846154,"corpus_reward":0.07692307692307693,"harness":"hermes","int_rate":0.5466666666666666,"intercepted":41,"matched":10,"model":"claude-opus-4-7","n":75,"partial":true,"reward":0.13333333333333333,"wall_h":null},{"corpus":"v2","corpus_intercepted":0.4846153846153846,"corpus_reward":0.18461538461538463,"harness":"hermes","int_rate":0.4846153846153846,"intercepted":63,"matched":24,"model":"glm-5.1","n":130,"reward":0.18461538461538463,"wall_h":null},{"corpus":"v2","corpus_intercepted":0.3,"corpus_reward":0.06923076923076923,"harness":"hermes","int_rate":0.48148148148148145,"intercepted":39,"matched":9,"model":"gpt-5.5","n":81,"partial":true,"reward":0.1111111111111111,"wall_h":null},{"corpus":"v2","corpus_intercepted":0.43846153846153846,"corpus_reward":0.1,"harness":"hermes","int_rate":0.43846153846153846,"intercepted":57,"matched":13,"model":"deepseek-v4-pro","n":130,"reward":0.1,"wall_h":null},{"corpus":"v2","corpus_intercepted":0.14615384615384616,"corpus_reward":0.046153846153846156,"harness":"hermes","int_rate":0.14615384615384616,"intercepted":19,"matched":6,"model":"openrouter-owl-alpha","n":130,"reward":0.046153846153846156,"wall_h":null},{"corpus":"v2","corpus_intercepted":0.03076923076923077,"corpus_reward":0.015384615384615385,"harness":"hermes","int_rate":0.03076923076923077,"intercepted":4,"matched":2,"model":"deepseek-v4-flash","n":130,"reward":0.015384615384615385,"wall_h":null}],"sort_key":"intercepted / corpus_size (V2: 130, V1: 153) DESC, reward DESC"}
