Generative AI has permeated numerous products and services, with varying degrees of success. Among these, AI-powered coding stands out as a particularly effective application. Both individual developers and large corporations are increasingly relying on generative tools to create and debug programs. In an exciting development, Google has announced that Gemini Code Assist is now available free of charge for indie developers.
Gemini Code Assist was initially launched late last year as an enterprise tool. The newly released version retains almost all of the original features. While other AI models like ChatGPT can assist with coding queries, Gemini Code Assist is uniquely designed to integrate seamlessly with the tools developers already use. By connecting Gemini Code Assist to your development environment, you can leverage the power of a large language model (LLM) without the need to switch between windows. This tool remains aware of your code and is ready to offer suggestions. Users can also engage in code-related discussions with the model, provided the language is public domain.
Originally priced at $45 per month per user, Gemini Code Assist is now available at no cost for individual developers. The free tier is generous, offering 180,000 code completions per month, which Google claims is sufficient even for prolific professional developers. This is a stark contrast to Microsoft's GitHub Copilot, which offers only 2,000 code completions and 50 Copilot chat messages per month. By comparison, Gemini Code Assist provides 90 times the completions of Copilot, according to Google's calculations.
The technology underpinning Gemini Code Assist remains unchanged from its enterprise-only days. Based on the latest Gemini 2.0 LLM, it has been optimized for coding. Google highlights the importance of a large context window in making AI coding useful for developers, with Code Assist capable of handling complex programming problems thanks to its 128,000 input token limit.
Gemini Code Assist integrates with several popular IDEs and platforms, including Firebase, Visual Studio, and GitHub. This positions it in direct competition with Microsoft's less generous AI coding tools. Despite investing heavily in data centers to run AI workloads, Google acknowledges that inference is not free. By rolling out Gemini Code Assist with high limits, Google aims to embed developers into its Gemini ecosystem, potentially steering them away from Microsoft and OpenAI. Long-term success will depend on these AI tools standing the test of time.
Google notes that the new Gemini Code Assist for individuals is still in a preview phase, meaning functionality may evolve. Some features from the enterprise version, such as productivity metrics and customized AI responses, are absent. For those interested in these advanced features, the paid version remains an option.