changelog

Vercel AI SDK Adds Grok 4.7 Model ID Support in xai Provider Patch 3.0.134

TL;DR

Vercel released version 3.0.134 of its @ai-sdk/xai provider package, adding support for a Grok 4.7 model ID. The patch also bumps two internal dependencies, provider-utils and openai-compatible.

2 min read
0

Vercel shipped version 3.0.134 of @ai-sdk/xai, the xAI provider package for its AI SDK, adding a model ID reference for "Grok 4.7." The change appears in a single commit labeled feat(xai): add grok 4.7 model ID, part of a routine patch release.

The update also pulls in two dependency bumps: @ai-sdk/provider-utils to version 4.0.52 and @ai-sdk/openai-compatible to version 2.0.76. No other functional changes are listed in the release notes.

What this actually is

This is a patch to Vercel's AI SDK, the open-source toolkit developers use to integrate multiple model providers into JavaScript and TypeScript applications through a unified API. The xai provider package translates AI SDK calls into requests compatible with xAI's API. Adding a model ID string to this package lets developers reference "Grok 4.7" in their code when calling xAI's endpoints through the SDK.

Importantly, this release does not constitute an official model launch from xAI. There is no accompanying announcement from xAI confirming Grok 4.7's specifications, benchmark scores, context window, pricing, or release date. Model IDs are frequently added to SDK packages ahead of or around a provider's own announcement, sometimes even speculatively, to reduce integration lag once a model becomes available via API.

No details on Grok 4.7's parameter count, training cutoff, context window, or pricing per million tokens are available in this changelog or from xAI directly as of this release. Developers attempting to call the model ID through the SDK would need xAI's API to actually support that identifier server-side for requests to succeed.

What this means

For developers using the Vercel AI SDK, this patch means the string identifier for Grok 4.7 is now recognized by the xai provider package, which simplifies future integration once xAI formally exposes the model via its API. It does not confirm that Grok 4.7 is publicly available, fully trained, or benchmarked.

Teams should treat this as an infrastructure update rather than a signal to build production features around Grok 4.7 today. The authoritative source for Grok 4.7's capabilities, pricing, and availability will be xAI's own announcement, not a dependency bump in a third-party SDK. Until xAI publishes specifications, any claims about the model's performance or feature set remain unconfirmed.

Comments

Loading...