← All comparisons

Claude 4.5 Haiku (Reasoning) vs DeepSeek V4 Pro (Reasoning, Max Effort)

Anthropic vs DeepSeek — side-by-side benchmark comparison

Claude 4.5 Haiku (Reasoning)DeepSeek V4 Pro (Reasoning, Max Effort)
Intelligence Index37.151.5
Coding Index32.647.5
Math Index83.7
Output speed (tok/s)142.251.8
Blended price ($/1M)$2.19$0.54
Time to first token (s)10.48s1.21s
aime
aime 2583.7%
artificial analysis coding index32.6047.50
artificial analysis intelligence index37.1051.50
artificial analysis math index83.70
gpqa67.2%88.8%
hle9.7%35.9%
ifbench54.3%76.5%
lcr70.3%66.3%
livecodebench61.5%
math 500
mmlu pro76.0%
scicode43.3%50.0%
tau254.7%96.2%
terminalbench hard27.3%46.2%

Benchmark data from Artificial Analysis.