Vercel released version 4.0.58 of its @ai-sdk/xai package on September 11 at 18:59, according to a GitHub changelog entry. This update includes new functionality for handling image generation requests in batches and implements fixes for specific issues.
Key Points
- The @ai-sdk/xai package was updated to version 4.0.58.
- The release occurred on September 11 at 18:59.
- The update adds support for image generation requests in batches.
- It includes a fix to reject unsupported batch request types.
- A specific fix addresses the preservation of reasoning streams across empty tool-call deltas for DeepSeek.
- The @ai-sdk/provider dependency was updated to version 4.0.14.
- The @ai-sdk/provider-utils dependency was updated to version 5.0.40.
Context
The @ai-sdk/xai package is part of the Vercel AI SDK, a collection of tools designed to help developers build AI-powered applications. This update, identified by commit 9ed46d2, focuses on enhancing the SDK's capabilities for handling image generation and refining its interaction with specific models like DeepSeek.
Why It Matters
This update indicates Vercel's ongoing development of its AI SDK, particularly in expanding its multimodal capabilities and improving reliability for developers. The addition of batch image generation support can streamline workflows for applications requiring multiple image outputs, while the DeepSeek fix suggests attention to model-specific integration details.
What To Do
- Developers using the Vercel AI SDK should review the changelog for @ai-sdk/xai@4.0.58.
- Note the new support for batch image generation if building applications that require this functionality.
- If using DeepSeek with the SDK, observe the improved handling of reasoning streams.
- Compare the updated dependency versions for @ai-sdk/provider and @ai-sdk/provider-utils with current project configurations.
