Google has made a significant leap forward in artificial intelligence (AI) with the release of Gemini 2.0, a model laden with autonomous capabilities and multimodal features. With this advancement, Google’s vision of AI chatbots as evolving AI agents comes to fruition. These agents are customized software that utilize generative AI for real-time interactions and task execution with users.
The new model enhances Gemini 1.5’s multimodal foundations by adding native image generation and text-to-speech capabilities, along with improved reasoning skills. Google claims that the 2.0 Flash variant outperforms the previous 1.5 Pro model on key benchmarks while operating at twice the speed.
The model is currently accessible to users who subscribe to Google Advanced, a paid service competing against Claude and ChatGPT Plus. For those willing to dive deeper, the model can be accessed via Google AI Studio. Despite its complex interface, it offers a more comprehensive experience, allowing users to upload up to 1 million tokens of context, nearly 10 times ChatGPT’s capacity, and control the model’s creativity or factuality through adjustable settings.
Google’s new model also introduces a “Deep Research” feature. This feature utilizes the model’s enhanced reasoning and long-context capabilities to explore complex topics and compile reports, presenting a plan that users can edit and instruct the chatbot to execute. This level of control and customization is unmatched by any AI service currently on the market.
Google concurrently shared a video showcasing Project Astra, an experimental AI assistant powered by Gemini 2.0. Project Astra is equipped with expanded capabilities such as Multilingual conversations with improved accent recognition, integration with Google Search, Lens, and Maps, an extended memory that retains 10 minutes of conversation context, long-term memory, and low conversation latency through new streaming capabilities.
While the reception on social media has been lukewarm, the new models seem to be gaining traction among users, especially in light of a major blackout of ChatGPT Plus.
Indeed, Google’s announcement comes amid OpenAI’s “12 Days of Christmas” campaign, where the company unveils a new product daily. Google also unveiled Project Mariner, an AI-powered Chrome extension that uses agents to navigate websites and complete tasks.
Google plans to integrate Gemini 2.0 throughout its product lineup, starting with experimental access to the Gemini app, with a broader release following in January. This includes integration into Google Search’s AI features, which currently reach over 1 billion users.
Meanwhile, Anthropic quietly released its latest update, Claude 3.5 Haiku, a faster version of its AI models. While still in training, the model is already demonstrating superior performance on coding tasks. Both Google’s and Anthropic’s premium services are priced at $20 monthly, matching OpenAI’s basic ChatGPT Plus tier.
Anthropic’s Claude 3.5 Haiku is proving to be faster, cheaper, and more potent than previous models, scoring highly on HumanEval coding tasks and multilingual math problems. The company claims users can achieve up to 90% cost savings through prompt caching and an additional 50% reduction using the Message Batches API. This positions the model as a cost-effective option for enterprises looking to scale their AI operations and provides stiff competition for OpenAI’s offerings.
