The latest addition to the family of AI models, , is now generally available for production usage. Google revealed on Thursday that the former was a smaller and faster model than the Gemini 1.5 Flash which was produced at Google I/O. Due to its pace, it has very low latency inference and is more efficient in the generation of output. Importantly, the tech leader shared that Flash-8B AI model is the “lowest cost per intelligence of any Gemini model”.
Gemini 1.5 Flash-8B AI Model Now Generally AvailableIn a recent developer blog post, the Mountain View-based tech company spelled out the new AI model. The company had previously distilled the Gemini 1.5 Flash AI model into the Gemini 1.5 Flash-8B, which was targeted towards more efficient generation and faster processing of the output. Now, it’s claiming that Google DeepMind developed this smaller version of the AI model a few months ago.
While the new model is smaller in size, the tech giant states that the latter “nearly matches” the 1.5 Flash model in multiple benchmarks, ranging from simple chat to transcription and long context language translation.
The biggest advantage of the AI model is that it is priced effectively. Google stated that the Gemini 1.5 Flash-8B will come with the lowest token price in the Gemini family. Developers are going to have to pay $0.15 (about Rs. 12.5) per one million output tokens, $0.0375 (about Rs. 3) per one million input tokens and $0.01 (about Rs. 0.8) for each one million tokens on cached prompts.
Google is also doubling the rate limits for the 1.5 Flash-8B AI model. Developers can now send up to 4,000 requests per minute (RPM) while making use of this model. As the tech giant explained, the model has been chosen because it proves handy for simple, high-volume tasks. Developers who would like to try out the model can do so free of cost through Google AI Studio and Gemini API.
