← All comparisons

Claude 4.5 Haiku (Reasoning) vs Gemma 3 4B Instruct

Anthropic vs Google — side-by-side benchmark comparison

Claude 4.5 Haiku (Reasoning)Gemma 3 4B Instruct
Intelligence Index37.16.3
Coding Index32.62.9
Math Index83.712.7
Output speed (tok/s)142.20.0
Blended price ($/1M)$2.19$0.05
Time to first token (s)10.48s0.00s
aime6.3%
aime 2583.7%12.7%
artificial analysis coding index32.602.90
artificial analysis intelligence index37.106.30
artificial analysis math index83.7012.70
gpqa67.2%29.1%
hle9.7%5.2%
ifbench54.3%28.3%
lcr70.3%5.7%
livecodebench61.5%11.2%
math 50076.6%
mmlu pro76.0%41.7%
scicode43.3%7.3%
tau254.7%5.0%
terminalbench hard27.3%0.8%

Benchmark data from Artificial Analysis.