Gemini is the LLM offering from Google Deepmind. They claim that it outperforms GPT-4 in 30 out of 32 benchmarks tested, but a recent study (“An In-depth Look at Gemini’s Language Abilities”), claims otherwise, where Gemini underperforms GPT-3.5 for most tasks. Sources Gemini Landing Page Introducing Gemini: our largest and most capable AI model Google Gemini is not even as good as GPT-3.5 Turbo, researchers find An In-depth Look at Gemini’s Language Abilities