← All comparisons

Qwen3.5 0.8B (Non-reasoning) vs Devstral Small (Jul '25)

Alibaba vs Mistral — side-by-side benchmark comparison

Qwen3.5 0.8B (Non-reasoning)Devstral Small (Jul '25)
Intelligence Index9.915.2
Coding Index1.012.1
Math Index29.3
Output speed (tok/s)96.3183.4
Blended price ($/1M)$0.02$0.15
Time to first token (s)0.26s0.40s
aime0.3%
aime 2529.3%
artificial analysis coding index100.0%12.10
artificial analysis intelligence index9.9015.20
artificial analysis math index29.30
gpqa23.6%41.4%
hle4.9%3.7%
ifbench21.6%34.6%
lcr6.7%17.0%
livecodebench25.4%
math 50063.5%
mmlu pro62.2%
scicode2.9%24.3%
tau265.2%28.4%
terminalbench hard0.0%6.1%

Benchmark data from Artificial Analysis.