Open-source AI models are increasingly challenging proprietary solutions by offering more cost-effective alternatives. Developers and enterprises now have access to models from companies such as DeepSeek, Alibaba, Meta, Zhipu AI, and Tencent, which can be self-hosted or deployed via affordable inference providers.
Data indicates that the cost of inference for open-source models is 50–90% lower than that of proprietary AI APIs. This price gap is prompting some organizations, including Harvey, to transition away from established providers like OpenAI and Anthropic in favor of open-source alternatives.
The shift in pricing strategies is forcing major industry players to reconsider their market positions. Future performance benchmarks and potential updates from companies like Anthropic, Google, and Meta will likely play a critical role in determining how the competitive landscape evolves in response to these open-source developments.
