Google is preparing to launch its next multimodal AI model, Gemini 3, during its livestream event on October 9, aiming to rival OpenAI's GPT-5, reports stated. The model will have improvements in reasoning, coding accuracy, and creative generation. Another report said that Gemini 3 has outpaced its competitor, Anthropic's Claude 4.5, in handling vector graphics and structured outputs.