← All comparisons

Claude Sonnet 4.6 (Non-reasoning, High Effort) vs Claude 4.5 Haiku (Reasoning)

Anthropic vs Anthropic — side-by-side benchmark comparison

Claude Sonnet 4.6 (Non-reasoning, High Effort)Claude 4.5 Haiku (Reasoning)
Intelligence Index44.437.1
Coding Index46.432.6
Math Index83.7
Output speed (tok/s)55.2142.2
Blended price ($/1M)$6.56$2.19
Time to first token (s)1.07s10.48s
aime
aime 2583.7%
artificial analysis coding index46.4032.60
artificial analysis intelligence index44.4037.10
artificial analysis math index83.70
gpqa79.9%67.2%
hle13.2%9.7%
ifbench41.2%54.3%
lcr57.7%70.3%
livecodebench61.5%
math 500
mmlu pro76.0%
scicode46.9%43.3%
tau279.5%54.7%
terminalbench hard46.2%27.3%

Benchmark data from Artificial Analysis.