Agentic evaluation with extracted-text documents in a sandbox and code-execution tools; public Messages API; production safety classifiers active with fallback to Opus 4.8 on refusal; card-wide default max effort applies.
未知项
The exact run count is not stated in this section.
工具
允许工具
推理预算
max
框架
未记录
协议
OfficeQA Pro; 133-question subset; agentic extracted-text evaluation with code-execution tools.
条件
Public Messages API; safety classifiers and Opus 4.8 fallback active; max effort by card-wide default.
System Card max effort (per-result benchmark harnesses)
设置
System Card §8.1 default: adaptive thinking/max effort and default sampling settings unless a section says otherwise. Benchmark tools, evaluator, safeguards, fallback, and repetitions remain per-result conditions; this config represents the stated model effort, not harness equivalence.
未知项
Exact run time is not disclosed.、Per-benchmark tools, safeguards, and repetition counts are specified only where the source states them.
工具
允许工具
推理预算
max
框架
未记录
协议
OfficeQA Pro 133-question subset; max effort; mean over five runs; extracted Treasury Bulletin text in sandbox with code execution.
条件
Remain separate from full-set OfficeQA. Source-wide conditions: Extracted-text corpus in a sandbox; code execution tools; OfficeQA Pro is the 133-question subset.
Differences are calculated only for unique, active, exact values between direct predecessors when the unit, metric, and explicitly recorded configuration and conditions match.
· 1 score record for this release, including 1 exact value (may include historical records)
Records for the direct predecessor do not provide matching conditions for a valid difference.
Claude Opus 5 · OfficeQA scores and configurations
Metric and score
Configuration and evaluation conditions
Sources
Exact-match accuracy66.9%Reported 66.9% · Unit %
Published Unconfirmed
OfficeQA / OfficeQA Pro · Messages API
Settings
Agentic evaluation with extracted-text documents in a sandbox and code-execution tools; public Messages API; production safety classifiers active with fallback to Opus 4.8 on refusal; card-wide default max effort applies.
Unknowns
The exact run count is not stated in this section.
Tools
Tools allowed
Reasoning budget
max
Framework
not recorded
Protocol
OfficeQA Pro; 133-question subset; agentic extracted-text evaluation with code-execution tools.
Conditions
Public Messages API; safety classifiers and Opus 4.8 fallback active; max effort by card-wide default.
· 1 score record for this release, including 1 exact value (may include historical records)
Records for the direct predecessor do not provide matching conditions for a valid difference.
Claude Opus 5.5 · OfficeQA scores and configurations
Metric and score
Configuration and evaluation conditions
Sources
Exact-match accuracy67.7%Reported 67.7% · Unit %
Published Unconfirmed
System Card max effort (per-result benchmark harnesses)
Settings
System Card §8.1 default: adaptive thinking/max effort and default sampling settings unless a section says otherwise. Benchmark tools, evaluator, safeguards, fallback, and repetitions remain per-result conditions; this config represents the stated model effort, not harness equivalence.
Unknowns
Exact run time is not disclosed.、Per-benchmark tools, safeguards, and repetition counts are specified only where the source states them.
Tools
Tools allowed
Reasoning budget
max
Framework
not recorded
Protocol
OfficeQA Pro 133-question subset; max effort; mean over five runs; extracted Treasury Bulletin text in sandbox with code execution.
Conditions
Remain separate from full-set OfficeQA. Source-wide conditions: Extracted-text corpus in a sandbox; code execution tools; OfficeQA Pro is the 133-question subset.