← All comparisons

Claude Sonnet 4.6 (Non-reasoning, Low Effort) vs DeepSeek V3.2 Exp (Reasoning)

Anthropic vs DeepSeek — side-by-side benchmark comparison

Claude Sonnet 4.6 (Non-reasoning, Low Effort)DeepSeek V3.2 Exp (Reasoning)
Intelligence Index42.632.9
Coding Index43.033.3
Math Index87.7
Output speed (tok/s)54.90.0
Blended price ($/1M)$6.56$0.31
Time to first token (s)1.13s0.00s
aime
aime 2587.7%
artificial analysis coding index43.0033.30
artificial analysis intelligence index42.6032.90
artificial analysis math index87.70
gpqa79.7%79.7%
hle10.8%13.8%
ifbench42.4%54.1%
lcr58.7%69.0%
livecodebench78.9%
math 500
mmlu pro85.0%
scicode44.1%37.7%
tau278.9%33.9%
terminalbench hard42.4%31.1%

Benchmark data from Artificial Analysis.