Home Model series Claude Haiku Claude Haiku 4.5 Claude Haiku 4.5 · OSWorld-Verified · Verified benchmark; evaluation protocol noted per source · Scores and configurations Claude Haiku 4.5 · OSWorld-Verified · Verified benchmark; evaluation protocol noted per source · Scores and configurations Claude Haiku 4.5 record for OSWorld-Verified · Verified benchmark; evaluation protocol noted per source. 1 score record for this release, including 1 exact value (may include historical records). Includes 1 source location.
Scores and evaluation conditions Claude Haiku 4.5 · OSWorld-Verified scores and configurations Metric and score Configuration and evaluation conditions Sources First-attempt success rate 50.7% Reported 50.7% · Unit % Published Unconfirmed
发布表报告值 Settings SWE-bench: full 500 problems, 50 trials, no test-time compute, 128K thinking budget, default sampling; Terminal-Bench: Terminus 2, 11 runs split between no thinking and 32K thinking; OSWorld: OSWorld-Verified, 100 steps, 4 runs; GPQA/MMMLU/MMMU follow source table. Unknowns AIME 的 Python 工具使用情况按两个分列配置保存;其余各项的逐行配置以主文脚注为准。 Tools Tool access not recorded Reasoning budget Thinking budget 128K total thinking budget and 2K thinking budget per-step Framework OSWorld-Verified Protocol 官方 OSWorld-Verified framework;100 steps;4 runs。 Conditions 128K total thinking budget, 2K per step。 View this score in rankings
Benchmark versions and metrics Aliases OSWorld Verified
First-attempt success rate · % · Higher is better