← All comparisons

Claude 4.5 Haiku (Non-reasoning) vs Phi-3 Mini Instruct 3.8B

Anthropic vs Microsoft — side-by-side benchmark comparison

Claude 4.5 Haiku (Non-reasoning)Phi-3 Mini Instruct 3.8B
Intelligence Index31.010.1
Coding Index29.63.0
Math Index39.00.3
Output speed (tok/s)105.80.0
Blended price ($/1M)$2.19$0.00
Time to first token (s)0.65s0.00s
aime4.0%
aime 2539.0%0.3%
artificial analysis coding index29.603.00
artificial analysis intelligence index31.0010.10
artificial analysis math index39.0030.0%
gpqa64.6%31.9%
hle4.3%4.4%
ifbench42.0%23.9%
lcr43.7%2.0%
livecodebench51.1%11.6%
math 50045.7%
mmlu pro80.0%43.5%
scicode34.4%9.0%
tau232.5%0.0%
terminalbench hard27.3%0.0%

Benchmark data from Artificial Analysis.