← All comparisons

Claude 4.5 Haiku (Non-reasoning) vs DeepSeek V3 0324

Anthropic vs DeepSeek — side-by-side benchmark comparison

Claude 4.5 Haiku (Non-reasoning)DeepSeek V3 0324
Intelligence Index31.022.3
Coding Index29.622.0
Math Index39.041.0
Output speed (tok/s)105.80.0
Blended price ($/1M)$2.19$1.21
Time to first token (s)0.65s0.00s
aime52.0%
aime 2539.0%41.0%
artificial analysis coding index29.6022.00
artificial analysis intelligence index31.0022.30
artificial analysis math index39.0041.00
gpqa64.6%65.5%
hle4.3%5.2%
ifbench42.0%41.0%
lcr43.7%41.0%
livecodebench51.1%40.5%
math 50094.2%
mmlu pro80.0%81.9%
scicode34.4%35.8%
tau232.5%47.1%
terminalbench hard27.3%15.2%

Benchmark data from Artificial Analysis.