← All comparisons

Claude 4.5 Haiku (Non-reasoning) vs MiniMax-M2.5

Anthropic vs MiniMax — side-by-side benchmark comparison

Claude 4.5 Haiku (Non-reasoning)MiniMax-M2.5
Intelligence Index31.041.9
Coding Index29.637.4
Math Index39.0
Output speed (tok/s)105.8190.8
Blended price ($/1M)$2.19$0.53
Time to first token (s)0.65s2.38s
aime
aime 2539.0%
artificial analysis coding index29.6037.40
artificial analysis intelligence index31.0041.90
artificial analysis math index39.00
gpqa64.6%84.8%
hle4.3%19.1%
ifbench42.0%71.6%
lcr43.7%66.0%
livecodebench51.1%
math 500
mmlu pro80.0%
scicode34.4%42.6%
tau232.5%95.3%
terminalbench hard27.3%34.8%

Benchmark data from Artificial Analysis.