Models/Claude Opus 5.5

Claude Opus 5.5 openrouter/anthropic/claude-opus-5.5@anthropic Anthropic effort low

85%

85 of 100 right on its latest run, Sep 25. Its first verdict comes on Oct 2, 2026.

Runs so far1

Verdict fromOct 2

Baseline so far: Sep 25, 2026 (1 run, 100 answers).

Score, every run

Share of questions answered correctly. Grey band: the baseline's normal range. Hover for a run.

70%80%90%100%Sep 25

All runs so farAnswersRight

Reasoning20100%

Logic puzzles2070%

Reading code2095%

Following instructions2090%

Long documents2070%

Per runAverage so farLatest run

Refused to answer0.00.0

Input tokens per question6,5996,599

Output tokens per question2,0542,054

Thinking tokens per question1,3761,376

Cost of 100 questions$3.43$3.43