Home Model series Claude Opus Claude Opus 4.5 Claude Opus 4.5 · AA-Omniscience · Anthropic factuality evaluation on 42 economically relevant topics across six domains; Claude Sonnet 4.5 grader; dataset revision not stated · Scores and configurations Claude Opus 4.5 · AA-Omniscience · Anthropic factuality evaluation on 42 economically relevant topics across six domains; Claude Sonnet 4.5 grader; dataset revision not stated · Scores and configurations Claude Opus 4.5 record for AA-Omniscience · Anthropic factuality evaluation on 42 economically relevant topics across six domains; Claude Sonnet 4.5 grader; dataset revision not stated. 6 score records for this release, including 6 exact values (may include historical records). Includes 1 source location.
Scores and evaluation conditions Claude Opus 4.5 · AA-Omniscience scores and configurations Metric and score Configuration and evaluation conditions Sources Correct rate 43.7% Reported 43.7% · Unit % Published Unconfirmed
Public factuality datasets, no thinking Settings No thinking; no external tools stated; Claude Sonnet 4.5 grades Simple-QA-Verified and AA-Omniscience. Unknowns Trial count and model sampling settings are not stated. Tools Tool access not recorded Reasoning budget not recorded Framework not recorded Protocol Anthropic evaluation on AA-Omniscience; correct rate; Sonnet 4.5 grader. Conditions No thinking; trial count and sampling settings not stated. View this score in rankings Incorrect rate 36.7% Reported 36.7% · Unit % Published Unconfirmed
Public factuality datasets, no thinking Settings No thinking; no external tools stated; Claude Sonnet 4.5 grades Simple-QA-Verified and AA-Omniscience. Unknowns Trial count and model sampling settings are not stated. Tools Tool access not recorded Reasoning budget not recorded Framework not recorded Protocol Anthropic evaluation on AA-Omniscience; incorrect rate; Sonnet 4.5 grader. Conditions No thinking; trial count and sampling settings not stated. View this score in rankings Uncertain response rate 19.6% Reported 19.6% · Unit % Published Unconfirmed
Public factuality datasets, no thinking Settings No thinking; no external tools stated; Claude Sonnet 4.5 grades Simple-QA-Verified and AA-Omniscience. Unknowns Trial count and model sampling settings are not stated. Tools Tool access not recorded Reasoning budget not recorded Framework not recorded Protocol Anthropic evaluation on AA-Omniscience; uncertain response rate; Sonnet 4.5 grader. Conditions No thinking; trial count and sampling settings not stated. View this score in rankings Correct rate 51.9% Reported 51.9% · Unit % Published Unconfirmed
Public factuality datasets, 16K extended thinking Settings 16K extended-thinking budget; no external tools stated; Claude Sonnet 4.5 grades Simple-QA-Verified and AA-Omniscience. Unknowns Trial count and model sampling settings are not stated. Tools Tool access not recorded Reasoning budget Thinking budget 16K Framework not recorded Protocol Anthropic evaluation on AA-Omniscience; correct rate; Sonnet 4.5 grader. Conditions 16K extended-thinking budget; trial count and sampling settings not stated. View this score in rankings Incorrect rate 36.6% Reported 36.6% · Unit % Published Unconfirmed
Public factuality datasets, 16K extended thinking Settings 16K extended-thinking budget; no external tools stated; Claude Sonnet 4.5 grades Simple-QA-Verified and AA-Omniscience. Unknowns Trial count and model sampling settings are not stated. Tools Tool access not recorded Reasoning budget Thinking budget 16K Framework not recorded Protocol Anthropic evaluation on AA-Omniscience; incorrect rate; Sonnet 4.5 grader. Conditions 16K extended-thinking budget; trial count and sampling settings not stated. View this score in rankings Uncertain response rate 11.5% Reported 11.5% · Unit % Published Unconfirmed
Public factuality datasets, 16K extended thinking Settings 16K extended-thinking budget; no external tools stated; Claude Sonnet 4.5 grades Simple-QA-Verified and AA-Omniscience. Unknowns Trial count and model sampling settings are not stated. Tools Tool access not recorded Reasoning budget Thinking budget 16K Framework not recorded Protocol Anthropic evaluation on AA-Omniscience; uncertain response rate; Sonnet 4.5 grader. Conditions 16K extended-thinking budget; trial count and sampling settings not stated. View this score in rankings
Benchmark versions and metrics Correct rate · % · Higher is better Incorrect rate · % · Lower is better Uncertain response rate · % · Lower is better