Microsoft's Copilot Cowork switches to usage-based billing, considers DeepSeek model
Microsoft moves Copilot Cowork to usage-based billing, mulls cheaper DeepSeek model.

Microsoft is considering a fine-tuned version of DeepSeek V4 as a cheaper model option for Copilot Cowork. The company is also switching to usage-based billing, as Copilot head Charles Lamanna says flat-rate pricing isn't sustainable. This move aligns with a broader industry pattern.
Microsoft's potential adoption of DeepSeek V4 for Copilot Cowork comes as the company looks to optimize costs. A fine-tuned version of the model could offer a more affordable alternative for users. The shift to usage-based billing reflects the evolving needs of the market.
According to Charles Lamanna, flat-rate pricing models are no longer viable. This change may impact how businesses and developers utilize Copilot Cowork, potentially leading to more flexible and scalable solutions. The integration of a potentially cheaper model and a usage-based billing system could make Copilot Cowork more attractive to a wider range of users.
This strategic move positions Microsoft to better compete in the AI-powered coworking space. Why this matters: The transition to usage-based billing and the potential incorporation of DeepSeek V4 into Copilot Cowork have significant implications for the AI and developer communities. By adopting a more flexible pricing model, Microsoft is acknowledging the diverse needs of its users and the fluctuating demands of AI-powered applications.
This shift may prompt other industry players to reevaluate their pricing strategies, leading to a more dynamic and user-centric market. For developers and businesses, the availability of a potentially cheaper, fine-tuned model like DeepSeek V4 could enable more cost-effective solutions and broader adoption of AI-powered coworking tools. However, questions remain about the performance and reliability of the DeepSeek model in real-world applications and how Microsoft will balance the benefits of usage-based billing with the need for predictable costs and scalability.
Source: The Decoder