Is the Price War of Large Language Models Coming? Google's Gemini1.5Flash Update

TapTechNews August 3rd news, is the price war of large language models coming? Google updated its price page yesterday (August 2nd) and announced that starting from August 12th, 2024, the cost of each million input tokens of the Gemini1.5Flash model will be $0.075 and the cost of each million output tokens will be $0.3 (TapTechNews note: currently about 2.2 Chinese yuan).

This makes the usage cost of the Gemini1.5Flash model nearly 50% cheaper than that of OpenAI's GPT-4omini. According to the calculation, the input cost of the Gemini1.5Flash model has decreased by 78.6% compared to before, and the output cost has decreased by 71% compared to before.

TapTechNews compared the cost situations in a chart as follows:

Cost per million input tokens Cost per million output tokens Gemini1.5 Flash Current $0.35 $1.05 Gemini1.5 Flash New $0.075 $0.3 OpenAI GPT-4o mini $0.15 $0.6

In terms of performance, the Gemini1.5Flash is still lagging behind GPT-4omini. As shown in the following table, in all top AI benchmark tests except MathVista, GPT-4omini is superior to Gemini1.5Flash.