Vercel has updated its AI SDK, specifically the @ai-sdk/prodia package to version 1.0.54, to include a warning mechanism. This change, released on August 26, addresses the provision of an unsupported seed language model setting within the Prodia integration.
Key Points
- The @ai-sdk/prodia package was updated to version 1.0.54.
- This update introduces a warning for unsupported seed language model settings.
- The change was released on August 26.
- The @ai-sdk/zai package, version 2.0.0, added the Z.AI provider with GLM chat completions, streaming, reasoning, tools, and multimodal inputs.
- The Z.AI provider and AI Gateway now support the GLM-5.3-Flash model.
- The @ai-sdk/xai@3.0.127 package preserves complete xAI Responses and Chat Completions usage objects in raw usage metadata and generated/streamed results.
- The @ai-sdk/amazon-bedrock@4.0.163 package preserves complete Amazon Bedrock Converse usage objects in raw usage metadata.
Context
According to the Vercel AI SDK Changelog, the @ai-sdk/prodia update focuses on improving developer feedback by flagging unsupported configurations. Concurrently, broader updates across the AI SDK include new provider integrations and enhanced usage data preservation. For instance, the @ai-sdk/zai package now integrates the Z.AI provider, offering capabilities such as GLM chat completions and multimodal inputs, and has added support for the GLM-5.3-Flash model to both the Z.AI provider and the AI Gateway.
Why It Matters
These updates indicate Vercel's ongoing efforts to refine developer experience and expand the capabilities of its AI SDK. Builders can benefit from clearer guidance on model settings and access to new providers and models, which can impact the range of AI applications they can develop and deploy.
What To Do
- Note the warning for unsupported seed language model settings when using @ai-sdk/prodia.
- Explore the new Z.AI provider and its features, including GLM-5.3-Flash model support, if building with the Vercel AI SDK.
- Review the updated usage object preservation in @ai-sdk/xai and @ai-sdk/amazon-bedrock for more complete raw usage metadata.
