Google Gemini 1.5: The Next Generation AI with Advanced Language Learning Abilities
Google has just unveiled its latest AI model, Gemini 1.5, which is set to revolutionize the field of artificial intelligence. This new version comes just a few months after the initial Gemini reveal and promises to be a mind-blowing upgrade that will put tremendous pressure on competitors like OpenAI and ChatGPT.
Gemini 1.5 Pro, the free version that most people will get, is already as good as the high-end Gemini 1.0 Ultra model, according to Google. But what sets Gemini 1.5 apart is its ability to support a context window of up to 1 million tokens. This means that users can input massive documents and even large videos while interacting with the chatbot for information. Furthermore, Gemini 1.5 can learn from context, allowing it to teach itself a rare language after reading a grammar book, similar to how a person would learn a language but at a much faster pace.
Google’s demos of Gemini have significantly improved since its initial launch. The new demos showcase the capabilities of interacting with Gemini and give users a glimpse of what to expect. Google also mentions whether the videos are sped up to replicate real-life conditions.
The most significant upgrade in Gemini 1.5 Pro is its support for 1 million tokens, which allows the AI to process over 700,000 words, over 30,000 lines of code, 11 hours of audio, or 1 hour of video content. Google reveals that during their research, they even went up to 10 million tokens, although such an upgrade may only come much later down the road, if ever.
When the free Gemini 1.5 Pro version launches, it will initially support about an eighth of the 1 million tokens, or 128,000 tokens. The paid versions will go up to 1 million tokens. Developers and enterprise customers, on the other hand, will have access to the best version (1 million tokens) while Gemini 1.5 Pro is in testing.
Gemini 1.5 also boasts a new Mixture-of-Experts (MoE) architecture, making it more efficient. This architecture allows Gemini to use only parts of its Large Language Model (LLM) to answer queries, resulting in faster performance for the end user. Additionally, Google provides several examples of what Gemini 1.5 Pro can do, such as finding funny exchanges in the 402-page transcripts from Apollo 11’s mission to the moon or reproducing the text from a silent Buster Keaton movie.
Gemini 1.5 Pro’s “in-context learning” skills enable it to learn a new skill from the information provided in a long prompt without fine-tuning. For example, when given a grammar manual for a language with fewer than 200 speakers worldwide, Gemini 1.5 Pro can learn to translate English to that language at a similar level to a person learning from the same content.
Developers and enterprises can start testing Gemini 1.5 Pro right now through Google’s AI Studio and Vertex AI products. The public rollout for Gemini 1.5 Pro (128,000 tokens) will follow soon, and Google will announce pricing tiers for users who want even more context, up to 1 million tokens.
With this groundbreaking upgrade, Google’s Gemini 1.5 is set to challenge competitors like OpenAI and ChatGPT. It will be interesting to see how these companies respond to Google’s latest move in the AI space.
