Maca2024

Cache hit rate

97%

Input served from cache

Higher than 76% of developers

Cache reuse

31.5×

Reads per cached token

Higher than 63% of developers

Cost per 1M tokens

$0.59

Blended cost across all tokens

Lower than 85% of developers

Output share

0%

Tokens generated as output

Higher than 82% of developers

Parallel tools

15%

Days with 2+ tools · 1.2 avg/day

Parallel machines

5%

Days with 2+ machines · 1.0 avg, 2 max

Models per active day

2.5

Distinct models used each day

Daily cost variation

1.3

Coefficient of variation; lower is steadier

Active days in span

26%

First to last active day