DeepSeek V4 Flash and V4 Pro Now Available in Microsoft Foundry

On May 8, 2026, Microsoft revealed the addition of DeepSeek V4 Flash to its Foundry model catalog, marking a significant expansion of enterprise AI options. DeepSeek V4 Flash is engineered for low latency and high throughput, specifically targeting cost-sensitive applications where speed and efficiency are critical. The model is available immediately, with the more powerful DeepSeek V4 Pro slated for a future release. This integration allows enterprises to deploy production-ready AI systems directly within Microsoft's cloud ecosystem, leveraging DeepSeek's optimized architecture for real-time inference and large-scale operations.

The move underscores Microsoft's strategy to offer diverse, specialized models alongside its own GPT-based solutions. By including DeepSeek's cost-efficient Flash variant, Microsoft aims to attract enterprises seeking to balance performance against operational budgets. The upcoming V4 Pro is expected to deliver even greater capabilities for complex tasks. This partnership signals DeepSeek's growing influence in the market, competing with models like Meta's Llama and Anthropic's Claude on the Azure platform, giving businesses more flexibility in choosing the right AI for their workloads.