Claude Opus 4.1 simulates the user; Anthropic added a prompt addendum; corrected task setup and grading.
未知项
Effort, trial count and detailed sampling settings are not stated.
工具
工具条件未记录
推理预算
未记录
框架
未记录
协议
τ²-bench Airline corrected variant; Table 2.8.A.
条件
Claude Opus 4.1 simulated the user; Anthropic added a task prompt addendum. The card states that task/setup/grading corrections were submitted to the benchmark authors and warns that Airline is unsuitable for cross-model comparisons or as a reliable policy-adherence measure.
Claude Opus 4.5 · τ²-bench · Anthropic corrected Airline variant; task and grading corrections submitted to evaluation authors · Scores and configurations
Claude Opus 4.5 record for τ²-bench · Anthropic corrected Airline variant; task and grading corrections submitted to evaluation authors. 1 score record for this release, including 1 exact value (may include historical records). Includes 1 source location.
Claude Opus 4.5 · τ²-bench scores and configurations
Metric and score
Configuration and evaluation conditions
Sources
Category score87.8%Reported 87.8% · Unit %
Published Unconfirmed
τ²-bench Airline corrected variant
Settings
Claude Opus 4.1 simulates the user; Anthropic added a prompt addendum; corrected task setup and grading.
Unknowns
Effort, trial count and detailed sampling settings are not stated.
Tools
Tool access not recorded
Reasoning budget
not recorded
Framework
not recorded
Protocol
τ²-bench Airline corrected variant; Table 2.8.A.
Conditions
Claude Opus 4.1 simulated the user; Anthropic added a task prompt addendum. The card states that task/setup/grading corrections were submitted to the benchmark authors and warns that Airline is unsuitable for cross-model comparisons or as a reliable policy-adherence measure.