Large Language ModelIs the price war coming?GoogleThe Company updated its pricing page to announce that beginning August 12, 2024, theGemini The 1.5 Flash model costs $0.075 per million input tokens and $0.3 per million output tokens (currently about $2.2).
This makes the Gemini 1.5 Flash model nearly 50% cheaper to use than OpenAI's GPT-4o mini.Based on the calculations, the Gemini 1.5 Flash model cost input cost is 78.61 TP3T lower than before and the output cost is 711 TP3T lower than before.
A graphical comparison of costs is shown below:
|
|
Cost per million input tokens |
Cost per million output tokens |
|
Gemini 1.5 Flash Current |
0.35 USD |
1.05 dollars |
|
Gemini 1.5 Flash New |
0.075 dollars |
0.3 USD |
|
OpenAI GPT-4o mini |
0.15 USD |
0.6 USD |
In terms of performance, the Gemini 1.5 Flash still lags behind the GPT-4o mini, as shown in the table below, which outperforms the Gemini 1.5 Flash in all of the top AI benchmarks except MathVista.
