Mozilla Report Finds Open AI Models Are Nearly Catching Up With Big Tech.

Mozilla’s inaugural State of Open Source AI report highlights a major shift in the artificial-intelligence industry: open models are no longer simply experimental alternatives to systems developed by companies such as OpenAI, Google and Anthropic. According to Mozilla’s research, open AI models have moved much closer to the performance of leading closed models, changing the competitive landscape for developers and businesses.

One of the report’s most important findings is the shrinking capability gap. Mozilla’s research suggests that open models have reached a point where they can compete with proprietary systems on many important tasks. This development is particularly significant because open models can generally be downloaded, modified or deployed with fewer restrictions than fully closed commercial systems. The result is a growing opportunity for developers, researchers and smaller companies that cannot afford the enormous infrastructure required to build frontier AI models from scratch.

However, Mozilla also identifies a major imbalance between usage and revenue. Reports summarising the study indicate that open-source AI models account for roughly a third of AI usage while capturing only around 4% of the industry’s revenue. That suggests that the technology is becoming widely adopted without generating financial returns comparable to those enjoyed by companies operating proprietary AI platforms.

Another challenge is getting open AI projects from experimentation into reliable production systems. Mozilla’s research points to deployment, governance and operational tooling as continuing obstacles, with many open-source AI projects failing to make the transition from development to real-world production. This means that having a powerful model available publicly is only part of the equation; organizations also need infrastructure, security, monitoring and governance to use it effectively.

The broader implications are significant for the AI industry. If open models continue improving while remaining cheaper and more accessible, the advantage held by large proprietary AI companies could become increasingly difficult to maintain. Recent reporting also shows that Chinese and other open-model developers are narrowing the gap with Silicon Valley’s frontier systems, intensifying competition around model performance, price and accessibility.