Claude Opus 5.5 · ProgramBench(166 个保留的 golden tasks;架构与时限按协议区分) 91.2% · 附件证据(ev-op55-card-programbench)
Claude Opus 5.5 的 ProgramBench · 166 个保留的 golden tasks;架构与时限按协议区分 官方出处。System Card max effort (per-result benchmark harnesses) · 91.2%。见《Claude Opus 5.5 System Card》§8.10.1, pp.183-184。证据 ev-op55-card-programbench。
Claude Opus 5.5 · ProgramBench(166 个保留的 golden tasks;架构与时限按协议区分),System Card max effort (per-result benchmark harnesses)配置,Hidden test pass rate 91.2%(具体成绩公布时间未确认)。见《Claude Opus 5.5 System Card》§8.10.1, pp.183-184。 查看成绩条件
We excluded 34 tasks ... leaving 166 tasks ... using the same mini-swe-agent harness the upstream repo uses, without the six-hour time limit. On this set, Claude Opus 5.5 scored 91.2%.
官方当前 System Card;封面日期 2026-09-22;本地 PDF/TXT:本地留存副本、current-opus-5-5-system-card.txt;230 页全文核读;当前修订时间未知
版本公布
未确认
披露时间
未确认 · 首发/后补未确认
采集时间
历史依据
暂无可靠历史版本依据
引用文本
Claude Opus 5.5 · ProgramBench(166 个保留的 golden tasks;架构与时限按协议区分),System Card max effort (per-result benchmark harnesses)配置,Hidden test pass rate 91.2%(具体成绩公布时间未确认)。官方出处 Claude Opus 5.5 System Card(附件,官方当前 System Card;封面日期 2026-09-22;本地 PDF/TXT:本地留存副本、current-opus-5-5-system-card.txt;230 页全文核读;当前修订时间未知)§8.10.1, pp.183-184,https://www.anthropic.com/claude-opus-5-5-system-card#page=183。Benchlog https://benchlog.lekshome.cc/evidence/claude-opus-5-5/ev-op55-card-programbench/
关联成绩与条件
Claude Opus 5.5 · ProgramBench 成绩与配置
指标与成绩
配置与评测条件
证据关系
Hidden test pass rate91.2%原文 91.2% · 单位 %
公布 未确认
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
工具条件未记录
推理预算
max
框架
mini-SWE-agent harness
协议
Single-agent ProgramBench; 166 retained golden tasks; mini-swe-agent harness; no six-hour time limit; no internet or decompilation.
条件
The four-task difference from the multi-agent section is not merged; score uses only tests passed by each reference binary. Attempt count is not stated. Source-wide conditions: No internet or decompilation; score only tests passed by the reference binary; do not apply the 91.2% to all 200 tasks.
Claude Opus 5.5 · ProgramBench (166 个保留的 golden tasks;架构与时限按协议区分) 91.2% · Attachment evidence (ev-op55-card-programbench)
Official source for ProgramBench · 166 个保留的 golden tasks;架构与时限按协议区分 in Claude Opus 5.5. System Card max effort (per-result benchmark harnesses) · 91.2%. See “Claude Opus 5.5 System Card” at §8.10.1, pp.183-184. Evidence ev-op55-card-programbench.
Claude Opus 5.5 · ProgramBench (166 个保留的 golden tasks;架构与时限按协议区分), configuration System Card max effort (per-result benchmark harnesses), Hidden test pass rate 91.2% (exact score publication date unconfirmed). See “Claude Opus 5.5 System Card” at §8.10.1, pp.183-184. View score conditions
We excluded 34 tasks ... leaving 166 tasks ... using the same mini-swe-agent harness the upstream repo uses, without the six-hour time limit. On this set, Claude Opus 5.5 scored 91.2%.
官方当前 System Card;封面日期 2026-09-22;本地 PDF/TXT:本地留存副本、current-opus-5-5-system-card.txt;230 页全文核读;当前修订时间未知
Version published
Unconfirmed
Disclosure date
Unconfirmed · Launch or later disclosure unconfirmed
Collected
Historical basis
No reliable historical version evidence available
Citation text
Claude Opus 5.5 · ProgramBench (166 个保留的 golden tasks;架构与时限按协议区分), configuration System Card max effort (per-result benchmark harnesses), Hidden test pass rate 91.2% (exact score publication date unconfirmed). Official source: Claude Opus 5.5 System Card (Attachment, 官方当前 System Card;封面日期 2026-09-22;本地 PDF/TXT:本地留存副本、current-opus-5-5-system-card.txt;230 页全文核读;当前修订时间未知), §8.10.1, pp.183-184, https://www.anthropic.com/claude-opus-5-5-system-card#page=183. Benchlog https://benchlog.lekshome.cc/evidence/claude-opus-5-5/ev-op55-card-programbench/?lang=en
Related scores and conditions
Claude Opus 5.5 · ProgramBench scores and configurations
Metric and score
Configuration and evaluation conditions
Evidence relation
Hidden test pass rate91.2%Reported 91.2% · 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
Tool access not recorded
Reasoning budget
max
Framework
mini-SWE-agent harness
Protocol
Single-agent ProgramBench; 166 retained golden tasks; mini-swe-agent harness; no six-hour time limit; no internet or decompilation.
Conditions
The four-task difference from the multi-agent section is not merged; score uses only tests passed by each reference binary. Attempt count is not stated. Source-wide conditions: No internet or decompilation; score only tests passed by the reference binary; do not apply the 91.2% to all 200 tasks.