Grok 4.0 Goes GA in Microsoft Foundry and Grok 4.1 Fast Arrives with Major Enhancements

We first brought Grok 4.0 to Microsoft Foundry in September 2025, marking an important milestone in expanding Foundry’s multi-model ecosystem with frontier reasoning models from xAI. Since then, customer interest and usage have continued to build as developers explored Grok’s strengths in fast reasoning, sense-making, and interpretation of complex, ambiguous information. Today, we’re excited to announce that Grok 4.0 is now generally available (GA) in Microsoft Foundry, giving enterprises a production-ready path to deploy Grok at scale. 

Building on that momentum, Grok 4.1 Fast Non Reasoning , is now available in Microsoft Foundry and Grok 4.1 Fast Reasoning is coming soon. Grok 4.1 introduces a suite of improvements that enhance conversation quality, creativity, and emotional intelligence while maintaining core reasoning strengths. According to xAI, Grok 4.1 delivers more natural, fluid dialogue compared with earlier versions.  

Introducing Grok 4.1 Fast (Reasoning and Non-Reasoning) 

Grok 4.1 Fast is optimized for speed, scale, and agentic execution, giving developers flexibility to choose between reasoning and non-reasoning variants depending on workload requirements. 

• Grok 4.1 Fast (Reasoning): 
  Designed for scenarios that require rapid multi-step reasoning, structured decision-making, and interpretation of complex inputs. This variant is well-suited for agent workflows, analysis pipelines, and applications that need fast responses without sacrificing reasoning depth. 

• Grok 4.1 Fast (Non-Reasoning): 
  Optimized for maximum throughput and low latency, this variant is ideal for tasks such as summarization, classification, content transformation, and tool-driven execution where deterministic speed and efficiency matter more than deep reasoning. 

Together, these options allow teams to right-size performance and cost by selecting the appropriate Grok 4.1 Fast variant for each stage of an application from high-volume preprocessing and orchestration to targeted reasoning tasks. 

What’s New with Grok 4.1 Fast? 

Grok 4.1 brings several enhancements that broaden the model’s applicability and user experience: 

• Improved Conversational Quality: According to xAI, Grok 4.1 Fast offers smoother, more natural interaction patterns, making it more comfortable and intuitive to engage with, especially in multi-turn dialogues. 

• Enhanced Creativity and Emotional Awareness: According to xAI, Grok 4.1 Fast demonstrates stronger creative writing capabilities and greater emotional intelligence, helping it generate more expressive and engaging outputs that better align with human expectations.  

• Reduced Hallucination and Better Reliability: According to xAI, Grok 4.1 Fast produces fewer factual inaccuracies than its predecessor 

These enhancements make Grok 4.1 Fast a compelling choice for use cases that require engaging conversational interfaces, creative support, and rich natural language interaction. 

As with all frontier AI models, Grok-4.1 Fast introduces new capabilities alongside new operational considerations. Microsoft’s safety and responsible AI evaluations indicate that Grok-4.1 Fast may demonstrate increased risks in safety testing compared with other models available through Azure. In practice, this means there may be an increased risk of generating explicit or potentially harmful content. 

To support responsible deployment, customers should implement system-level safety instructions and leverage Azure AI Content Safety (AACS) to help monitor and filter outputs. Because no single safety system can address every possible risk scenario, customers are encouraged to conduct their own evaluations and validation before deploying Grok-4.1 in production systems. 

To provide enhanced safety and enterprise reliability, Microsoft's deployment of Grok 4.1 features a system-applied safety prompt that cannot be disabled. Customers are expected to operate the model without attempting to bypass or interfere with this feature. 

Enterprise-Ready Deployment via Azure AI Foundry 

With Grok 4.0 now GA in Foundry, enterprises gain the ability to incorporate advanced reasoning models into their workflows while enjoying the governance, compliance, and operational tooling that Azure provides.  

Models hosted in Foundry can be deployed serverless or with provisioned throughput, and customers benefit from centralized billing, identity integration, and access to other Azure services. 

Foundry’s model catalog also includes other Grok variants such as Grok 4.1 Fast and related non-reasoning SKUs, giving enterprises flexibility to balance performance, latency, and cost depending on their workloads.  

Pricing 

Looking Ahead 

The combination of Grok’s deep reasoning capabilities with the enterprise readiness of Microsoft Foundry opens new possibilities for production AI applications, from complex analytical agents and research assistants to creative and customer-facing experiences.  

With Grok 4.1’s conversational refinements further raising the model’s usability and expressiveness, Foundry customers can now experiment with and scale a broader set of AI-driven solutions, all within a trusted, governed environment. 

As Microsoft continues to expand Foundry’s catalog and partners like xAI continue to innovate, organizations have more options than ever to power next-generation AI applications across industries, use cases, and domains.  

Try Grok 4.1 Non-Reasoning  <AI Model Catalog | Microsoft Foundry Models>