Home Model series Claude Opus Claude Opus · ARC-AGI-1 · ARC Prize Foundation private dataset · Release history Claude Opus · ARC-AGI-1 · ARC Prize Foundation private dataset · Release history View scores, configurations, and disclosures for ARC-AGI-1 · ARC Prize Foundation private dataset across 2 Claude Opus releases, ordered by release date.
Differences are calculated only for unique, active, exact values between direct predecessors when the unit, metric, and explicitly recorded configuration and conditions match.
2025-11-24 · 1 score record for this release, including 1 exact value (may include historical records)
Claude Opus 4.5 · ARC-AGI-1 scores and configurations Metric and score Configuration and evaluation conditions Sources Accuracy 80% Reported 80.0% · Unit % Published Unconfirmed
ARC Prize Foundation private validation, 64K thinking Settings 64K thinking tokens; ARC Prize Foundation private validation dataset. Unknowns Run count and evaluation harness details are not stated in the card. Tools Tool access not recorded Reasoning budget Thinking budget 64K Framework not recorded Protocol ARC-AGI-1 private validation dataset; ARC Prize Foundation reported result. Conditions 64K thinking tokens; the system card says the model was trained on the public ARC-AGI-1 dataset. View this score in rankings
2026-04-16 · 2 score records for this release, including 2 exact values (may include historical records)
Claude Opus 4.7 · ARC-AGI-1 scores and configurations Metric and score Configuration and evaluation conditions Sources Accuracy 92% Reported 92.0% · Unit % Published Unconfirmed
ARC-AGI-1,max effort Settings ARC Prize private validation set; max effort; standard sampling; score is separate from the High-effort run. Unknowns none recorded Tools Tool access not recorded Reasoning budget max Framework not recorded Protocol ARC-AGI-1 private validation; max effort. Conditions Summary Table 8.1.A; separate from the 93.5% High-effort result in §8.11. View this score in rankings Accuracy 93.5% Reported 93.5% · Unit % Published Unconfirmed
ARC-AGI-1,high effort Settings ARC Prize private validation set; high effort; score is separate from the Max-effort summary-table value. Unknowns none recorded Tools Tool access not recorded Reasoning budget high Framework not recorded Protocol ARC-AGI-1 private validation; high effort. Conditions §8.11 body text; higher than the Max-effort result, which the card attributes to context exhaustion at the highest thinking levels. View this score in rankings
Table 8.1.A, PDF p.192; §8.11, p.212 Attachment · Claude Opus 4.7 System Card