← All comparisons

Claude 4.5 Haiku (Reasoning) vs DeepSeek R1 0528 (May '25)

Anthropic vs DeepSeek — side-by-side benchmark comparison

Claude 4.5 Haiku (Reasoning)DeepSeek R1 0528 (May '25)
Intelligence Index37.127.1
Coding Index32.624.0
Math Index83.776.0
Output speed (tok/s)142.20.0
Blended price ($/1M)$2.19$2.06
Time to first token (s)10.48s0.00s
aime89.3%
aime 2583.7%76.0%
artificial analysis coding index32.6024.00
artificial analysis intelligence index37.1027.10
artificial analysis math index83.7076.00
gpqa67.2%81.3%
hle9.7%14.9%
ifbench54.3%39.6%
lcr70.3%54.7%
livecodebench61.5%77.0%
math 50098.3%
mmlu pro76.0%84.9%
scicode43.3%40.3%
tau254.7%36.5%
terminalbench hard27.3%15.9%

Benchmark data from Artificial Analysis.