BREAKING
Customers Cut AI Costs Over 90%
0
M
planned AI spend ($)
0
x
cheaper model used
0
%
cost cut achieved
Open Models 87% Cheaper
Closed
1.86
Open/alt
0.23
Three Ways to Cut the Bill
1
Model routing
↓
2
Prompt caching
↓
3
Open models
Providers Feel the Pressure
OpenAI
Jalapeño chip
●
Weighing steep price cuts
●
New inference chip with Broadcom
Anthropic
-67% Opus
●
Cut Opus pricing in cases
●
Caching saves up to 90%
Flagships Still Needed for Hard Tasks
AI NEWS BLITZ
Big Anthropic and OpenAI customers are slashing their AI bills by more than ninety percent.