← All comparisons

Claude 4.5 Haiku (Reasoning) vs Devstral Small 2

Anthropic vs Mistral — side-by-side benchmark comparison

Claude 4.5 Haiku (Reasoning)Devstral Small 2
Intelligence Index37.119.5
Coding Index32.620.7
Math Index83.734.3
Output speed (tok/s)142.275.2
Blended price ($/1M)$2.19$0.00
Time to first token (s)10.48s0.73s
aime
aime 2583.7%34.3%
artificial analysis coding index32.6020.70
artificial analysis intelligence index37.1019.50
artificial analysis math index83.7034.30
gpqa67.2%53.2%
hle9.7%3.4%
ifbench54.3%31.2%
lcr70.3%24.0%
livecodebench61.5%34.8%
math 500
mmlu pro76.0%67.8%
scicode43.3%28.8%
tau254.7%23.4%
terminalbench hard27.3%16.7%

Benchmark data from Artificial Analysis.