Vercel AI SDK adds Grok 4.3, reasoning effort controls, and image quality model support for xAI

Vercel released version 3.0.97 of its AI SDK's xAI integration on June 24, 2024, adding support for three new models and reasoning effort controls.

New models and capabilities

The update adds support for:

• Grok 4.3: xAI's latest text model
• Grok Build 0.1: A new specialized model (specific capabilities not disclosed)
• Grok Imagine Image Quality: An image generation model focused on quality

The integration now supports reasoning effort settings with two options: 'none' and 'medium'. This allows developers to control the computational intensity of model responses, similar to reasoning controls in other models like OpenAI's o1 series.

Technical details

The patch (version 3.0.97) curates the model lineup to match xAI's current offerings. The reasoning effort parameter enables developers to trade off between response speed and depth of reasoning for compatible models.

Pricing for the new models and specific benchmark performance metrics were not disclosed in the release notes.

Integration context

Vercel's AI SDK provides a unified interface for integrating multiple AI providers into applications. The xAI integration allows developers to use Grok models alongside other providers like OpenAI, Anthropic, and Google through a single SDK.

The release is commit ab415f6 with patch changes documented under ticket 9300b02.

What this means

This update positions Vercel's AI SDK as one of the first major developer frameworks to support xAI's expanding model lineup. The addition of reasoning effort controls suggests xAI may be implementing similar capabilities to OpenAI's reasoning models, though the company has not publicly disclosed details about these features. For developers using Vercel's framework, this provides immediate access to Grok 4.3 without waiting for official xAI SDK updates, though integration through third-party SDKs may lag behind features available in xAI's native API.