> For clean Markdown of any page, append .md to the page URL. > For a complete documentation index, see https://docs.vellum.ai/changelog/2025/2025-02/llms.txt. > For AI client integration (Claude Code, Cursor, etc.), connect to the MCP server at https://docs.vellum.ai/_mcp/server. # Changelog | February, 2025 ## Support for GPT-4.5 Preview via OpenAI **February 27th, 2025** We've added support for the [GPT-4.5 Preview and Snapshot 02/27/2025](https://platform.openai.com/docs/models/gpt-4-5#gpt-4-5) via OpenAI. ## Support for Llama 3.2 and Llama 3.3 models via AWS Bedrock **February 27th, 2025** We've added support for [Llama 3.2 and Llama 3.3 models](https://docs.aws.amazon.com/bedrock/latest/userguide/models-supported.html) via AWS Bedrock on the `us-west-2` region. We've added the following models: * Llama 3.2 1B Instruct * Llama 3.2 3B Instruct * Llama 3.2 11B Instruct * Llama 3.2 90B Instruct * Llama 3.3 70B Instruct ## Support for Claude 3.7 Sonnet via Anthropic **February 24th, 2025** We've added support for Anthropic's new Claude 3.7 Sonnet model (latest and 02/19/2025 snapshot variants). ![Claude 3.7 Sonnet](https://www.anthropic.com/news/claude-3-7-sonnet) ## Workflow SDK Code Preview **February 18th, 2025** For those participating in the Workflow SDK beta, you can now preview the SDK-code representation of your Workflows directly in the UI. To do so, click on the preview button: ![Workflow SDK Code Preview Button](https://storage.googleapis.com/vellum-public/help-docs/changelogs/2025-02/preview-sdk-code-button.png) Doing so will open a side panel with a full representation of your Workflow in code form. ![Workflow SDK Code Preview Side Panel](https://storage.googleapis.com/vellum-public/help-docs/changelogs/2025-02/preview-sdk-code-side-panel.png) This feature is still in beta, so please let us know if you encounter any issues or have feedback! You can learn more about how to opt into the Workflows SDK beta [here](https://docs.vellum.ai/changelog/2025/2025-01#beta-release-of-sdk-enabled-workflows). ## Restore Button Moved to History Cards **February 16th, 2025** The Restore button has been moved from the top-right of Sandboxes pages to the History Cards themselves. This makes it easier to find and use the Restore button and revert back to a prior state of your Prompt or Workflow Sandbox. ## Workflow Node Layout Improvements **February 16th, 2025** There has been a minor improvement to the layout of Workflow Nodes. Now, input variables are always displayed above output types, making it easier to skim and understand a Node from top to bottom. Code Execution, Templating, and Final Output Nodes are all affected by this change. ## Model Picker Improvements **February 16th, 2025** The Model Picker has been updated to make it easier to find the model you're looking for. Now, you can sort models alphabetically by their name, the date they were introduced, and when they were last used by someone in your Workspace. ![Model Picker Improvements](https://storage.googleapis.com/vellum-public/help-docs/changelogs/2025-02/model-picker-improvements.png) Additionally, filters and sort preferences are saved across sessions, so you don't have to reapply them every time. ## Multiple Chat History Variables in Prompts **February 16th, 2025** Up until now, we've restricted you to only have one Chat History variable in a Prompt. It's quite uncommon to need more than one, but we understand that there are some cases where it can be useful. We've now lifted this restriction and you can now have multiple Chat History variables in a Prompt. ## Application-Wide Performance Improvements **February 10th, 2025** We recently overhauled core pieces of our webserver to improve the performance of the Vellum web application across the board. You should generally notice snappier page load times and more responsive interactions throughout the app. ## Support for xAI as a Model Host and Grok Models **February 7th, 2025** We've added support for xAI as a Model Host and along with it, the following [Grok models](https://docs.x.ai/docs/models) have been implemented: * Grok Beta * Grok Vision Beta * Grok 2 1212 * Grok 2 Vision 1212 ## Support for Multiple Gemini Models via Vertex AI **February 7th, 2025** We've added support for several new Gemini models via Vertex AI: * [Gemini 2.0 Flash 001](https://cloud.google.com/vertex-ai/generative-ai/docs/learn/models#gemini-2.0-flash) * [Gemini 2.0 Flash Lite](https://cloud.google.com/vertex-ai/generative-ai/docs/learn/models#gemini-2.0-flash-lite) * [Gemini 2.0 Pro](https://cloud.google.com/vertex-ai/generative-ai/docs/learn/models#gemini-2.0-pro) * [Gemini 2.0 Flash Experimental](https://cloud.google.com/vertex-ai/generative-ai/docs/learn/models#gemini-2.0-flash) * [Gemini 1.5 Flash](https://cloud.google.com/vertex-ai/generative-ai/docs/learn/models#gemini-1.5-flash) ## Support for DeepSeek V3 via Fireworks AI **February 7th, 2025** We've added support for the [DeepSeek V3](https://fireworks.ai/models/fireworks/deepseek-v3) model via Fireworks AI. ## Support for Gemini 2.0 Flash 001 via Gemini **February 7th, 2025** We've added support for the [Gemini 2.0 Flash 001](https://ai.google.dev/gemini-api/docs/models/gemini#gemini-2.0-flash) model via Gemini. ## Support for OpenAI's o3-mini 2025-01-31 Snapshot via Azure **February 3rd, 2025** We've added support for OpenAI's [o3-mini 2025-01-31 snapshot](https://platform.openai.com/docs/models/#o3-mini) to be used as a self-hosted model via Azure.