← All comparisons

Qwen3.5 397B A17B (Non-reasoning) vs DeepSeek V3.2 Exp (Reasoning)

Alibaba vs DeepSeek — side-by-side benchmark comparison

Qwen3.5 397B A17B (Non-reasoning)DeepSeek V3.2 Exp (Reasoning)
Intelligence Index40.132.9
Coding Index37.433.3
Math Index87.7
Output speed (tok/s)53.50.0
Blended price ($/1M)$1.35$0.31
Time to first token (s)1.85s0.00s
aime
aime 2587.7%
artificial analysis coding index37.4033.30
artificial analysis intelligence index40.1032.90
artificial analysis math index87.70
gpqa86.1%79.7%
hle18.8%13.8%
ifbench51.6%54.1%
lcr58.0%69.0%
livecodebench78.9%
math 500
mmlu pro85.0%
scicode41.1%37.7%
tau283.9%33.9%
terminalbench hard35.6%31.1%

Benchmark data from Artificial Analysis.