Vercel AI SDK Adds Support for Google's Deep Research Models and Gemini Embedding-2
Vercel released version 5.0.0-canary.98 of its Google Vertex AI SDK, adding support for three new Google models: deep-research-max-preview-04-2026, deep-research-preview-04-2026, and gemini-embedding-2. The update enables developers to integrate Google's research-focused models and latest embedding model into applications using Vercel's AI SDK.
Vercel shipped version 5.0.0-canary.98 of its Google Vertex AI SDK package, adding support for three new Google models: deep-research-max-preview-04-2026, deep-research-preview-04-2026, and gemini-embedding-2.
The update applies to both the @ai-sdk/google-vertex and @ai-sdk/google packages, bringing the same model support to developers using either Google's Vertex AI platform or standard Google AI services.
New Models Added
The deep-research-max-preview-04-2026 and deep-research-preview-04-2026 models represent Google's research-focused model variants, though specific capabilities and pricing have not been disclosed. The "04-2026" designation in the model names suggests an April 2026 preview snapshot.
The gemini-embedding-2 model is Google's latest text embedding model, succeeding the previous generation. Embedding models convert text into numerical vectors for semantic search, retrieval-augmented generation (RAG), and similarity matching applications. Technical specifications and benchmark performance metrics have not been released.
Package Updates
The release bumps @ai-sdk/google-vertex to version 5.0.0-canary.98 and @ai-sdk/google to version 4.0.0-canary.75. Both packages maintain compatibility with Vercel's AI SDK framework, which provides unified interfaces for working with multiple AI providers.
Developers using the Vercel AI SDK can now access these models through the same API patterns used for other Google models, without requiring separate integration work.
What This Means
This SDK update primarily benefits developers building on Vercel's platform who want early access to Google's research models. The canary release status indicates these are preview versions intended for testing rather than production deployment. The lack of detailed documentation on model capabilities suggests Google is testing these variants with a limited developer audience before broader release. Organizations evaluating embedding model options now have access to Google's latest generation through a standardized SDK interface.
Related Articles
Google removes model picker for free Gemini app users, locking them to 'Auto' routing that defaults to Flash-Lite
Google has removed manual model selection for free Gemini app users, leaving 'Auto' as the only option. Most prompts will go to Flash-Lite, with harder ones routed to Flash or Pro, according to Google. New thinking levels (Low, Medium, High) are also rolling out to subscribers.
Vercel AI SDK Patch Adds Support for 'None' Reasoning Effort on GPT-6 Sol and Luna
Vercel released version 4.0.78 of @ai-sdk/openai, a patch that adds support for setting reasoningEffort to 'none' for GPT-6 Sol and Luna models. The update also introduces validation logic that warns on unsupported request-level effort updates and rejects invalid historical ones.
Vercel's AI SDK Adds Support for Unannounced OpenAI Model 'GPT-6.1 Sol'
Vercel released @ai-sdk/openai version 3.0.121, a patch that adds SDK support for a model string called 'GPT-6.1 Sol.' OpenAI has not publicly announced or confirmed this model, and no specs, pricing, or benchmarks are available.
Vercel's AI SDK Adds Support for Unreleased 'GPT-6.1 Sol' Model Identifier
Vercel's @ai-sdk/openai package version 2.0.132 quietly adds support for a model identifier called 'GPT-6.1 Sol.' OpenAI has not confirmed the model's existence, specs, or release date.
Comments
Loading...