Vercel has updated its AI SDK, releasing version ai@7.0.85 with several patch changes. These updates address how image generation costs are calculated and improve the handling of Anthropic batch requests. The changes were part of a broader release that included updates to various AI SDK packages, such as @ai-sdk/valibot@3.0.34, @ai-sdk/xai@4.0.50, @ai-sdk/voyage@2.0.34, and @ai-sdk/vue@4.0.85.
Key Points
- The AI SDK version ai@7.0.85 includes patch changes.
- Gateway image-generation costs across split requests are now summed correctly.
- The Anthropic provider preserves native message batch request counts in provider metadata.
- The Anthropic provider now supports the full language-model option surface in batch requests.
- Individual image generation calls are now exposed.
- Anthropic reasoning budgets are enabled for application inference profile ARNs in @ai-sdk/amazon-bedrock@5.0.68.
Context
According to the Vercel AI SDK Changelog, these updates were released via the Changesets release GitHub action. The changes aim to refine the functionality and cost tracking within the AI SDK, particularly for image generation and interactions with the Anthropic API. The @ai-sdk/amazon-bedrock package also received an update to enable Anthropic reasoning budgets.
Why It Matters
These updates provide developers using the Vercel AI SDK with more accurate cost reporting for image generation and enhanced control over Anthropic batch requests. The exposure of individual image generation calls and support for Anthropic reasoning budgets can lead to more precise resource management and potentially more efficient application development.
What To Do
- Review the Vercel AI SDK documentation for details on the ai@7.0.85 patch changes.
- Test image generation cost tracking in applications using the AI SDK Gateway.
- Note the updated support for Anthropic batch requests and language-model options.
- Developers using @ai-sdk/amazon-bedrock should verify the impact of enabled Anthropic reasoning budgets.