← All comparisons

Claude 4.5 Haiku (Reasoning) vs Qwen3 4B 2507 Instruct

Anthropic vs Alibaba — side-by-side benchmark comparison

Claude 4.5 Haiku (Reasoning)Qwen3 4B 2507 Instruct
Intelligence Index37.112.9
Coding Index32.69.0
Math Index83.752.3
Output speed (tok/s)142.20.0
Blended price ($/1M)$2.19$0.00
Time to first token (s)10.48s0.00s
aime
aime 2583.7%52.3%
artificial analysis coding index32.609.00
artificial analysis intelligence index37.1012.90
artificial analysis math index83.7052.30
gpqa67.2%51.7%
hle9.7%4.7%
ifbench54.3%33.5%
lcr70.3%7.3%
livecodebench61.5%37.7%
math 500
mmlu pro76.0%67.2%
scicode43.3%18.1%
tau254.7%26.6%
terminalbench hard27.3%4.5%

Benchmark data from Artificial Analysis.