Test: Claude Opus 5.5 is cheaper and faster than Opus 5, but not smarter

Anthropic released Claude Opus 5.5, cutting API pricing from $5/$25 to $4/$20 per million tokens and claiming 40% savings and 30% faster output. An independent test on three reasoning tasks found identical results to Opus 5, real speedup of about 11%, and a safety-filter refusal on the hardest problem.

• Opus 5.5 API pricing: $4 per million input and $20 per million output tokens, down from $5/$25
• On the logic grid Opus 5.5 cost $0.16 versus $0.27 for Opus 5, a 43% saving
• In the stone game Opus 5.5 used 28,740 tokens and $0.58 versus 74,981 tokens and $1.88
• Output speed was 103.4 tokens/s for Opus 5.5 versus 93.1 for Opus 5, about 11% faster, not 30%