The way the software is built may not be the same. Advances in large-scale linguistic models have made it known to AI researcher Andrej Karpathy which is spoken of to last year as “vibe coding,” or using natural language to generate code. In accordance with Inferyamore than 25 million apps were built with vibe code tools as of 2026.
Today, Microsoft’s Github Copilot, Claude Code, OpenAI Codex and Cursor are among the most widely used AI coding tools. According to the intelligence provider It passesGitHub Copilot is used by 40 percent of large companies, but is the primary tool for only 17 percent of developers. Claude Code leads the main tools share with 28 percent, followed by Cursor with 24 percent, GitHub Copilot with 17 percent, and OpenAI Codex with 11 percent.
While competitors like Anthropic focus on building powerful coding assistants, GitHub positions itself as an ecosystem. It provides a centralized, cloud-based platform where developers can work directly in their integrated development environments (IDEs) with real-time code suggestions. It also provides access to many parameter models, including Claude Code and Codex, alongside open-source low-weight models.
This platform-based approach has become a key differentiator, as building applications involves much more than writing code. “It’s not just generating code. It’s about creating software, and that requires you to do more than just code part of it,” Kyle Daigle, GitHub’s chief operating officer and chief marketing officer (CMO) for developers at Microsoft, told the Observer.
A 13-year GitHub veteran—dating back to before Microsoft’s 2018 acquisition—Daigle helped bring Copilot to market in preview in June 2021. He now leads developer-first strategy across GitHub and the broader Microsoft ecosystem, including product launch, marketing and developer relations.
Copilot gives developers and businesses “the advantage of not being locked into one model family or needing to change the interface you’re using,” Daigle said, pointing out. “default mode” as a separator. This factor defines the purpose of the work, estimates the complexity and chooses the most economical model to complete the work.
AI coding costs
Cost has become a key factor in AI coding, with a range of industry shocks emerging from high token usage. Most notably, Uber it reportedly spent its AI budget in just four months. Accordingly, Axios shared the story of an anonymous AI consultant who said their company lost half a billion dollars after failing to impose usage restrictions on Claude Code licenses for employees.
Across the industry, retailers love it GitHub again Anthropic they have tried for payment changes. GitHub, for example, has moved from request-based billing to usage-based billing. Under the former, each interaction counted as one premium claim unit; under the latter, the cost varies based on the model used and the number of tokens used.
Usage-based payment has drawn criticism from some developers for driving up costs. “Continued criticism of GitHub’s usage-based pricing and reliability underscores the broader problem of an infrastructure that hasn’t kept up with the speed of AI coding,” Martin Reynolds, CTO of software delivery platform Harness, whose role includes working with customers to develop a product roadmap, told the Observer by email.
From this perspective, the industry as a whole not only needs to provide access to powerful coding tools, but needs to do so at scale in an economical way. Initiatives like the Copilot model harness and multiple model options can help control costs, but there is still more to do to support builders in the future.
