Changelog
API Docs
Help Center

February 13th, 2025

API

App

Models

Features

Characters

Venice.ai Change Log - February 13th, 2025

Multi-Modal Models / Launch of Qwen 2.5 VL 72B for Pro Users

We’re thrilled to release the number one most requested feature from Venice users!

  • Qwen 2.5 VL 72B, our first multi-modal vision model operating on Venice is now live for our Venice Pro users in both the Venice app, and via the Venice API. This is a vision model that lets you interact with images through the LLM. It can describe what it sees in images, and design prompts that you can use with Venice’s image models. We look forward to sharing additional content on its capabilities over the coming weeks and we are excited to hear about how you’re using it.

Venice Sticker Factory

The Venice Sticker factory is now live for all registered Venice users. Generate a sticker pack from any image created on Venice!

Here's how it works:

  1. Generate a new image. We've found prompts like "create a die cut sticker of the moon", using prompt enhance, and then running the result on Flux without styles works well as a starting place. Any image can be printed, but for die cut stickers, a clear border needs to be returned.

  2. In the actions under the image, click new sticker icon:

  1. Upload your sticker and continue.

  1. If your sticker has a clear border, you can remove the background. This will allow the printer to die cut the edges. Note, stickers may be printed smaller than 3×3” with the backgrounds removed.

  1. Continue to the order form.

  2. Complete your order and enjoy your new stickers in 5-7 days.

  3. Orders and shipping information will show up in your Order History on the left hand navigation:


Characters

  • Created a search and filtering mechanism to make finding characters easier.

API Dashboard

  • Users who have not met the qualifying criteria for API access will be shown a “Disabled” message in the API Tier box on the API Dashboard. API access requires either a Pro account, Staked VVV or a positive USD balance.

API

  • /chat/completions - Added support for the parallel_tool_calls option.

  • Added two headers to the response:

    • x-venice-balance-vcu - This will show the users VCU balance BEFORE the request has been processed.

    • x-venice-balance-usd - This will show the user’s USD balance BEFORE the request has been processed. It will only show up if the user’s VCU balance is less than or equal to 0.

Models

  • Context limit on Qwen 2.5 Coder 32B has been adjusted down to model’s limit of 32,768