Microsoft has introduced that DeepSeek R1 is now accessible via Azure AI Foundry and GitHub.
DeepSeek is an open supply AI mannequin created in China that has been a matter of dialog over the previous week after claims of how the mannequin was seemingly skilled for a fraction of what U.S.-based AI corporations have spent coaching their fashions and the way it makes use of much less computing energy to run.
Gross sales of Nvidia’s excessive performing chips have been restricted in China, forcing DeepSeek to coach these fashions on the lower-performance H800 chips. This has different corporations contemplating the chance that high-end chips is probably not as needed for AI improvement as beforehand thought. The R1 mannequin competes on efficiency and capabilities with fashions from OpenAI, Meta, and Google, however with considerably decrease prices.
“As a part of Azure AI Foundry, DeepSeek R1 is accessible on a trusted, scalable, and enterprise-ready platform, enabling companies to seamlessly combine superior AI whereas assembly SLAs, safety, and accountable AI commitments—all backed by Microsoft’s reliability and innovation,” Asha Sharma, company vice chairman of Microsoft’s AI Platform, wrote in a weblog put up.
The addition to Microsoft’s Azure AI Foundry and GitHub will allow builders to experiment with DeepSeek R1 whereas utilizing Microsoft’s built-in mannequin analysis instruments to check outputs and benchmark efficiency.
Microsoft additionally says it has put the mannequin via crimson teaming and security evaluations, together with automated assessments of mannequin conduct and safety opinions to mitigate potential dangers. Azure AI Content material Security gives built-in content material filtering by default, with the choice to opt-out, and the Security Analysis System permits functions to be examined earlier than they’re deployed.
“These safeguards assist Azure AI Foundry present a safe, compliant, and accountable setting for enterprises to confidently deploy AI options,” Sharma wrote.