Google announces Gemini Flash, a high-speed, high-performance lightweight AI model that is 1/10th the price of Gemini Pro and has the same performance
Google announced the lightweight and high-performance AI model ' Gemini Flash ' at 'Google I/O 2024' held on Wednesday, May 15, 2024. Gemini Flash is available at one-tenth the price of Gemini Pro, and has shown performance comparable to Gemini Pro in benchmark tests.
Gemini Flash - Google DeepMind
Google Gemini updates: Flash 1.5, Gemma 2 and Project Astra
https://blog.google/technology/ai/google-gemini-update-flash-ai-assistant-io-2024/
Gemini Flash is the lightest model in the Gemini series that operates via API. Gemini Flash was developed with an emphasis on processing speed, and the average latency is said to be less than 1 second for general use by developers and enterprises. In addition, Gemini Flash's context window is 1 million tokens, and it can process huge amounts of data such as '1 hour of video', '11 hours of audio', and 'more than 30,000 lines of code'.
Below is a table listing the benchmark results for the Gemini series. Gemini Flash significantly outperforms Gemini 1.0 Pro and even Gemini 1.5 Pro in some tests.
Gemini Flash is available through the Gemini API and Vertex AI. The Gemini API is available for free, and paid plans with relaxed rate limits are also available. The paid plan costs $0.35 (approximately 55 yen) per million tokens for prompts of up to 128,000 tokens, and $0.70 (approximately 110 yen) per million tokens for prompts of more than 128,000 tokens. Both prices are one-tenth of Gemini Pro.
Gemini Flash paid plans will be available from Thursday, May 30, 2024.
At the same time as the announcement of Gemini Flash, an update to Gemini Pro was also announced. The details of the update for Gemini Pro can be found in the following article.
Google updates Gemini 1.5 Pro, expanding context window from 1 million tokens to 2 million tokens - GIGAZINE
Related Posts:
in Software, Posted by log1o_hf