Release Note #19: Just Dropped – Gemini 3, GPT-5.1, Branching, and 7 Video Models (Plus a Big Win!)

Just dropped: Gemini 3 Pro and GPT-5.1 are live on Euryka. We introduce Branching in Threads to organise your workflows, add Google Veo and Sora to our video suite, and celebrate our recent win as New Business of the Year.

It’s been a little while since our last update. For the past few weeks, the team and I have been heads-down working on a series of under-the-hood upgrades across the entire platform.

These were foundational changes—the kind you don’t always see immediately, but you definitely feel. They were necessary to give us the stability and speed we need to build faster for you.

With that groundwork done, we are back with a massive update. We have two brand new frontier models, a game-changing workflow feature for Threads, and a significant expansion of our video capabilities.

Here is what’s new in Euryka.

- 02 EKRN19 Blog GPT5.1

Just Dropped: Gemini 3 Pro & GPT-5.1

It has been a massive week for AI, and we moved fast to get the latest models into your hands immediately.

First, we have enabled the Gemini 3 Pro Preview. This includes access to its new Thinking model, which offers enhanced reasoning capabilities for complex tasks.

Second, we have added GPT-5.1. You now have access to both the Base model and the Thinking model, which just dropped last Friday.

These are brand new. We are still testing them ourselves and exploring exactly what they can do, but we figured we’d let them out in the open for the community to start using immediately. You can access both right now in the model selector. Let us know what you find—we’re super excited to have them on the platform.

- 03 EKRN19 Blog BranchingThreads

Branching in Threads

A feature many of you requested is finally here. You can now create a branch from any point in a conversation.

This is perfect for keeping your workspace tidy. Instead of one endless, messy chat where ideas get buried, you can split long conversations into separate pathways. It is ideal for:

  • Exploring different creative directions without losing your original prompt
  • Testing ideas side-by-side
  • Keeping experiments separate from your main workflow

It keeps your threads clean, structured, and far easier to revisit.

- 04 EKRN19 Blog 7VideoModels

Seven Video Models. One Orchestrated Workflow.

We have added Google’s Veo 3 and Veo 3.1, plus OpenAI’s Sora 2, bringing the total number of supported video models on Euryka to seven:

  • Kling
  • Luma
  • RunwayML
  • Runway Aleph
  • Veo 3.0
  • Veo 3.1
  • Sora 2

The real power here is orchestration. You aren’t locked into a single ecosystem. You can generate an image in Midjourney, Imagen, Ideogram, or Flux, and then bring that image into any of these video models.

Use our Remix features to set start and end frames and build the exact shot or motion you need. It is about using the best model for the job, unified by your creative direction.

(Note: Veo and Sora are premium models with significant compute costs. If you’d like access, message us on Intercom and we’ll walk you through the setup.)

- 08 EKRN19 Blog Prospector

Prospector (Beta): Your Account-Based Marketing Just Got Intelligent

We’re rolling out something rather special. Prospector is our new lead discovery
engine, and it’s designed to cut through the noise.

Here’s how it works:

Ask Prospector to “Find all Marketing Directors and CMOs at B2B SaaS companies
in EMEA with 50-500 employees who raised Series A or B funding in the last 18
months.” Watch as it searches thousands of pages, discovers contacts, and builds
a targeted list for you.

Then take it further:

  • Run Deep Research on individual prospects
  • Build personalised email sequences based on real insights
  • Manage your entire account-based marketing workflow in one place

This is still in beta, so we’re working closely with early users to refine the
experience. Want in? Drop us a note on Intercom and we’ll add you to the beta
programme.

- 06 EKRN19 Blog FastPreviews

Faster Image Generation & Model Upgrades

We have tuned the engines to make image generation inside Threads significantly quicker. You will now get fast previews when ideating to keep the conversation flowing, with the option to refine and upscale further in Imaginations when you are ready for the final polish.

Alongside the headlines above, we have updated our underlying models across the rest of the platform. We have rolled out the latest versions of DeepSeek and Grok, ensuring you have faster reasoning and better accuracy across your Docs and Sparks workflows.

- 05 ENRN19 Blog Awards

We Won! New Business of the Year 🏆

Finally, we have a moment of pride to share. Euryka has just been named New Business of the Year at the Waterford Business Awards.

Building a global AI centre of excellence right here in the Déise is a mission we care deeply about. This award is a fantastic validation of that vision, but it is also thanks to you—our community—for trusting us with your brand.

Thank you for being part of this journey.


Ready to try the new models? Log in to Euryka to get started, or book a demo if you are new here.

Share this post on
Euryka AI Logo
©2025 Euryka AI Limited. All Rights Reserved

Book a demo to see Euryka in action!

Please use a business email to be connected with our sales team
By submitting this form, you agree to Euryka AI's Terms of Service and Privacy Policy.
Can't wait? Get started with Euryka today with our week pass for $9.99
Custom Cursor