Home Model series Claude Opus Claude Opus 4.6 Claude Opus 4.6 · GraphWalks · parents;256K–1M 上下文;题数未注明 · Scores and configurations Claude Opus 4.6 · GraphWalks · parents;256K–1M 上下文;题数未注明 · Scores and configurations Claude Opus 4.6 record for GraphWalks · parents;256K–1M 上下文;题数未注明. 2 score records for this release, including 2 exact values (may include historical records). Includes 2 source locations.
Scores and evaluation conditions Claude Opus 4.6 · GraphWalks scores and configurations Metric and score Configuration and evaluation conditions Sources Parents F1 71.1% Reported 71.1 (64k) · Unit % Published Unconfirmed
GraphWalks Parents 1M · 64K · internal Settings GraphWalks Parents 1M full variant;64K extended thinking;内部设置支持完整 prompt、thinking 与 output。 Unknowns effort 未说明;完整 1M variant 不可由公开 API 复现。 Tools Tool access not recorded Reasoning budget Thinking budget 64K Framework not recorded Protocol GraphWalks Parents 1M full variant;F1 score。 Conditions 5-run average;full variant includes problems beyond 1M API context; 64K extended thinking; effort unknown。 Parents F1 72% Reported 72.0 (max) · Unit % Published Unconfirmed
GraphWalks Parents 1M · Max · internal Settings GraphWalks Parents 1M full variant;Max effort;内部设置支持完整 prompt、thinking 与 output。 Unknowns 完整 1M variant 不可由公开 API 复现。 Tools Tool access not recorded Reasoning budget max Framework not recorded Protocol GraphWalks Parents 1M full variant;F1 score。 Conditions 5-run average;full variant includes problems beyond 1M API context; internal setting; max effort。
Benchmark versions and metrics Aliases GraphWalks Parents 256K-1M F1、GraphWalks Parents 1M full variant、GraphWalks Parents 1M F1
Accuracy · % · Higher is better Parents F1 · % · Higher is better