Open-source AI gains favour among new developers globally

A survey by Stack Overflow has highlighted a growing trend towards the use of open-source AI among developers, especially those at the early stages of their careers.

The survey engaged over 1,000 developers from around the world and found that 82% reported some level of experience with open-source AI. This trend is notably strong among young developers aged 25-34, with 60% reporting significant engagement. Early-career developers, those with less than five years of experience, exhibited the greatest familiarity with open-source AI at 71%. This indicates its rising importance for the next generation of developers.

Several AI models, specifically GPT-4o, DeepSeek, and models from Gemini, were frequently recognised among the participants. Notably, two of the top five models identified are open-source, underscoring the increasing significance of these tools within the developer community. When considering developers not working with open-source large language models, top awareness is present for Nova Pro (Amazon), Jamba, QwQ, Phi-4, and Mistral Large 2, with the exception of Nova Pro, all being open-source.

The survey also detailed the models most familiar to developers based on their years of experience. Developers with 0-5 years of experience were particularly familiar with DeepSeek (an open-source model) and Google's proprietary models. In contrast, those with 5-15 years of experience showed more awareness of larger language models (LLMs), with GPT-4o (33%), DeepSeek V3 (31%), and o1 (31%) being particularly noted.

Respondents ranked various models, with Claude Sonnet and o1 achieving high normalised rankings despite not being widely known, while Gemini models received lower average rankings despite widespread awareness. This highlights that some less popular models are highly valued by those familiar with them.

The survey also examined trust levels in AI tools, revealing that 66% of respondents entrusted open-source AI for personal projects and learning purposes. Although proprietary AI also enjoys trust, with 52% indicating it for personal and learning use, it is less trusted for strategic or development work. Security poses a concern regarding open-source AI for 44% of respondents, but for 48%, security is not a significant issue.

Erin Yepis, Senior Analyst, Market Research and Insights at Stack Overflow, commented on the findings: "Open-source AI is rapidly becoming a cornerstone for developers, especially those early in their careers. The high trust in open-source models for personal and learning projects underscores the community's confidence in these tools. As the technology evolves, we expect its adoption to continue growing, driven by the next generation of developers.''

Overall, the results of the survey point towards an increasing reliance and trust in open-source AI among developers, particularly in personal and educational settings. The expectation is that as technology progresses, the adoption of open-source AI will persist, propelled by younger developers and those at the onset of their careers.