A GenAI Fund survey of 2,719 approved AI builders across 55 countries finds OpenAI’s ChatGPT and Anthropic’s Claude effectively tied in overall use, while experience level and job role produce a clearer split: newcomers lean toward ChatGPT, and engineers and enterprise teams lean toward Claude.
As reported by TNGlobal, ChatGPT was used by 77.9 percent of respondents, narrowly ahead of Claude at 76.6 percent. Google Gemini ranked third at 61.1 percent. Those figures are not market share: 81.7 percent of surveyed builders reported using more than one platform, so respondents could select multiple tools.
The report defines beginners as people with one to two years of AI-tool experience, while experts have at least two years building production-ready code and enterprise systems.
Among builders with under a year of experience, ChatGPT led Claude by 79.3 percent to 70.9 percent. Claude moved ahead in the one-to-two-year group, 79.2 percent to 78.2 percent, and led more clearly among people with six to 10 years of experience, 78.4 percent to 72.4 percent. At three to five years, the two were essentially tied; among respondents with more than 10 years of experience, both sat at 73.8 percent.
The role breakdown follows a similar pattern. ChatGPT led among students, with 84.1 percent reporting use, versus 76.4 percent for Claude. Claude led among engineers, 84.4 percent to 79.2 percent, and among enterprise teams, 77.7 percent to 74.2 percent.
For Microsoft-centric shops, the low standalone figure for Azure OpenAI — 5.2 percent — should not be read as a direct measure of enterprise deployment. The survey asks about platforms used by individual builders, not cloud spending, managed-service consumption, or organization-wide production workloads. A developer may use ChatGPT directly while their application calls a model through Azure OpenAI, or vice versa.
The sample is also heavily technical. Of 1,952 respondents who provided language data, 72.5 percent used Python, 43.6 percent used JavaScript or TypeScript, and 17.5 percent used Java. That makes the results more useful as a view of active AI builders than of ordinary office users or Windows PC owners.
For admins and development leads, the practical message is to plan for mixed-model workflows rather than assume a single assistant will satisfy both new users and experienced engineering teams.
As reported by TNGlobal, ChatGPT was used by 77.9 percent of respondents, narrowly ahead of Claude at 76.6 percent. Google Gemini ranked third at 61.1 percent. Those figures are not market share: 81.7 percent of surveyed builders reported using more than one platform, so respondents could select multiple tools.
Experience is the dividing line
The report defines beginners as people with one to two years of AI-tool experience, while experts have at least two years building production-ready code and enterprise systems.Among builders with under a year of experience, ChatGPT led Claude by 79.3 percent to 70.9 percent. Claude moved ahead in the one-to-two-year group, 79.2 percent to 78.2 percent, and led more clearly among people with six to 10 years of experience, 78.4 percent to 72.4 percent. At three to five years, the two were essentially tied; among respondents with more than 10 years of experience, both sat at 73.8 percent.
The role breakdown follows a similar pattern. ChatGPT led among students, with 84.1 percent reporting use, versus 76.4 percent for Claude. Claude led among engineers, 84.4 percent to 79.2 percent, and among enterprise teams, 77.7 percent to 74.2 percent.
A multi-platform developer stack
Gemini’s 61.1 percent overall result shows the market is not a clean two-vendor contest, particularly for builders who use a mixture of chat interfaces, APIs, cloud services, and open-weight models. The survey also recorded use of DeepSeek at 18.2 percent, Hugging Face at 12.8 percent, Qwen at 11.6 percent, and LangChain at 9.8 percent. Some 27.5 percent said they used at least one Chinese AI model.For Microsoft-centric shops, the low standalone figure for Azure OpenAI — 5.2 percent — should not be read as a direct measure of enterprise deployment. The survey asks about platforms used by individual builders, not cloud spending, managed-service consumption, or organization-wide production workloads. A developer may use ChatGPT directly while their application calls a model through Azure OpenAI, or vice versa.
The sample is also heavily technical. Of 1,952 respondents who provided language data, 72.5 percent used Python, 43.6 percent used JavaScript or TypeScript, and 17.5 percent used Java. That makes the results more useful as a view of active AI builders than of ordinary office users or Windows PC owners.
For admins and development leads, the practical message is to plan for mixed-model workflows rather than assume a single assistant will satisfy both new users and experienced engineering teams.
References
- Primary source: TNGlobal
Published: 2026-07-17T16:44:39+00:00