Gemma: Introducing new state-of-the-art open models

Gemma: Introducing new state-of-the-art open models

Google get in on the openly licensed LLM game: Gemma comes in two sizes, 2B and 7B, trained on 2 trillion and 6 trillion tokens respectively. The terms of use "permit responsible commercial usage". In the benchmarks it appears to compare favorably to Mistral and Llama 2.

Something that caught my eye in the terms: "Google may update Gemma from time to time, and you must make reasonable efforts to use the latest version …

ai art benchmarks commercial game gemma generativeai google homebrewllms llama llama 2 llm llms mistral openly open models something state terms tokens usage