Google announces Gemini 2.0, its most advanced AI model
Google launched Gemini 2.0 this Thursday (May 29, 2026), the latest version of its multimodal artificial intelligence model. The announcement was made by the company itself, highlighting the system's ability to simultaneously process text, images, audio, and video, with a context window of up to 2 million tokens.
Availability for developers
Gemini 2.0 is now available via API for developers on Google AI Studio and Vertex AI. The novelty promises to boost applications requiring understanding of multiple input modalities, such as long video analysis, audio transcription and summarization, and processing large volumes of documents.
Superior performance on benchmarks
According to Google, Gemini 2.0 shows significant improvements in logical reasoning and code generation, surpassing previous benchmark results of GPT-4 (from OpenAI) and Claude 4 (from Anthropic). The company did not disclose specific numbers but claims the new model sets a new performance standard for complex AI tasks.
Implications for the market
The launch comes amid intense competition in the artificial intelligence sector, with major companies vying for leadership in generative AI models. Gemini 2.0 consolidates Google's presence in this market, offering a powerful tool for developers seeking to integrate multimodal capabilities into their products.
Next steps
Google has not disclosed when Gemini 2.0 will be available to the general public or if there will be a free version. For now, access is restricted to developers through the company's platforms.