Home Model series Claude Opus Claude Opus 4.6 Claude Opus 4.6 · ARC-AGI-2 · Verified;题集组成未在发布表中说明 · Scores and configurations Claude Opus 4.6 · ARC-AGI-2 · Verified;题集组成未在发布表中说明 · Scores and configurations Claude Opus 4.6 record for ARC-AGI-2 · Verified;题集组成未在发布表中说明. 1 score record for this release, including 1 exact value (may include historical records). Includes 2 source locations.
Scores and evaluation conditions Claude Opus 4.6 · ARC-AGI-2 scores and configurations Metric and score Configuration and evaluation conditions Sources Accuracy 68.8% Reported 68.8% · Unit % Published Unconfirmed
Max effort Settings 仅在来源明确写明 max effort 时使用;token budget 与 effort 分别记录,不由 token 数推断。 Unknowns none recorded Tools Tool access not recorded Reasoning budget max · Thinking budget 120K Framework not recorded Protocol ARC-AGI-2 (Verified);与 ARC Prize Foundation private dataset 成绩分开。 Conditions 68.8% 来自 Table 2.3.A 的 Verified 行;该表注明 Opus 4.6 通常使用 adaptive thinking + max effort,5-run average 除非另注。120K thinking tokens 属于另一个 ARC Prize private-dataset High-effort 结果,不适用于本记录。 View this score in rankings
Benchmark versions and metrics Aliases ARC-AGI-2 (Verified)
Accuracy · % · Higher is better