Euryka adds Claude Opus 4.5, GPT 5.2, DeepSeek 3.2, and updated Grok models. New agency-focused Sparks templates, Prospector beta updates, and advanced prompting guides for Flux 2 and Gemini Banana image models.
The AI landscape shifted dramatically this week. OpenAI issued what industry insiders are calling a ‘Code Red’ in response to Google Gemini’s rapid advances. Anthropic launched their most powerful reasoning model yet. And DeepSeek, the unexpected dark horse, continues to deliver impressive results at a fraction of the cost.
For marketing teams juggling multiple AI subscriptions, this creates a familiar problem: which model should you use? Which subscription is worth keeping? How do you stay current when the leaderboard changes every week?
Here’s the good news: with Euryka’s AI orchestration platform, you don’t need to pick a side. You get access to all the leading models in one workspace, with your brand context automatically applied across every interaction. When a new model launches, we integrate it. When an existing model improves, you benefit immediately.
This release adds four major model updates, five new agency-focused templates, and continued progress on our Prospector lead generation engine. Let’s dive in.
We’ve integrated the latest flagship models from the leading AI labs, giving you more choice depending on whether you need speed, depth, cost efficiency, or reasoning power.
Anthropic’s newest flagship reasoning model represents a significant leap in how AI approaches complex, multi-step problems. Unlike standard language models that generate responses in a single pass, Claude Opus 4.5 Thinking uses extended reasoning chains to break down complicated queries, verify its own logic, and deliver more accurate, nuanced outputs.
What makes it special:
Best for: Strategy documents, technical briefs, complex research synthesis, scenario planning, and any task requiring deep analytical thinking.
Early testing with our beta clients shows particularly strong results when using Opus 4.5 for competitive analysis, campaign strategy development, and technical documentation where accuracy matters more than speed.
OpenAI’s latest release arrives as a direct response to increasing competition from Google and Anthropic. Internally dubbed their ‘Code Red’ model, GPT 5.2 delivers noticeable improvements in creative writing quality, instruction following, and multi-format content generation.
Key improvements:
Best for: Blog posts, social media content, email campaigns, ad copy, and any creative writing where brand voice consistency matters.
If you’ve been using GPT-4 or GPT-4.5 for content generation, try GPT 5.2 on your next brief. The quality jump is immediately noticeable, particularly for maintaining consistent tone across longer documents.
DeepSeek remains the surprise story of 2025. Originally dismissed as a cost-focused alternative, it’s evolved into a genuinely competitive model that punches well above its weight class. DeepSeek 3.2 continues this trajectory with improvements in reasoning, coding, and structured output generation.
Why it matters:
Best for: Technical documentation, structured content, process workflows, templated outputs, and high-volume generation where cost efficiency matters.
Many of our agency clients use DeepSeek for first-draft generation and structured content, then refine with GPT or Claude for final polish. It’s a smart workflow that balances quality and cost.
We’ve updated Grok to the newest version, bringing improvements in conversational AI, real-time information access, and personality-driven outputs. Grok maintains its distinctive, more casual tone whilst delivering increasingly reliable results.
What’s improved:
Best for: Brainstorming sessions, trend analysis, exploratory conversations, and creative ideation where you want a more conversational AI partner.
With seven language models now available in Threads (Claude, GPT, Gemini, DeepSeek, Grok, Mistral, and Llama), here’s a quick decision framework:
Remember: with Euryka’s orchestration layer, you can switch models mid-conversation or use different models for different stages of the same project. Your brand context flows through automatically, so you maintain consistency regardless of which model you choose.
We’ve expanded our Sparks library with five new templates designed specifically for agency teams managing multiple clients, complex projects, and formal documentation requirements.
These aren’t generic templates. They’re built from proven agency workflows, pre-wired to pull from your Brand Hub and project context, and designed to reduce the time you spend on repetitive documentation.
Generate comprehensive SOWs that clearly define project scope, deliverables, timelines, and commercial terms. The Spark pulls client information from your project context, applies your agency’s standard terms, and produces a professional document ready for your final review.
What it includes:
Time saved: What typically takes 2-3 hours of copying, pasting, and reformatting now takes 10 minutes of context entry and review.
Create recurring service agreements with clear monthly deliverables, SLAs, and renewal terms. Particularly useful for agencies managing multiple ongoing client relationships where consistency across agreements matters.
What it includes:
Pro tip: Wire this Spark to your agency’s Brand Hub with your standard terms, then create project-specific versions for each client. The base template stays consistent whilst client-specific details adapt automatically.
Standardise how your creative team receives project instructions. This Spark generates detailed internal briefs that give designers and developers everything they need without endless Slack threads or unclear requirements.
What it includes:
Why it matters: Clear internal briefs reduce revision cycles, improve first-draft quality, and make your creative team more efficient. Many agencies report 30-40% fewer internal revisions after standardising their brief process.
Create client-facing creative briefs that clearly communicate project vision, requirements, and expectations. These briefs serve as alignment documents that reduce scope creep and ensure everyone agrees on objectives before work begins.
What it includes:
Best practice: Generate the external brief from your internal brief, then refine for client presentation. This ensures internal and external documentation stays aligned whilst tailoring language for each audience.
Transform completed projects into compelling case studies that demonstrate your agency’s capabilities and results. This Spark structures the narrative, highlights key metrics, and creates scannable content that works across your website, proposals, and sales conversations.
What it includes:
Time saved: Case studies that typically take a week of back-and-forth now come together in a day, giving you more marketing assets to support new business development.
These Sparks work best when you invest time upfront setting up your Brand Hub and project contexts properly. The more context you provide, the better the outputs become.
We’ve made significant progress on Prospector based on feedback from our early beta clients. Lots of UI enhancements, workflow improvements, and under-the-hood updates that make the entire lead discovery and outreach process feel more intuitive.
Improved UI/UX:
Enhanced Research Quality:
Workflow Improvements:
Account-based marketing typically requires three separate tools:
Each tool costs money. Each tool requires context switching. And your brand voice gets lost in the transitions.
Prospector consolidates this workflow into a single, brand-aware system:
Find: Query based on role, company size, industry, location, funding stage
Research: Deep Research on individual prospects to gather real insights
Personalise: Generate messaging sequences that reference specific company context
Launch: Manage outreach campaigns with Brand Hub tone automatically applied
Everything stays in one workspace. Your brand context flows through automatically. And you’re not copying and pasting between five different browser tabs.
Early beta clients are reporting:
One agency client recently told us: “Prospector turned our ABM process from a three-week slog into a three-day sprint. The quality didn’t drop. It improved.“
We’re working closely with a small group of beta clients to refine Prospector before wider release. If you’re interested in joining, reach out via Intercom and tell us:
We’re particularly interested in agencies running ABM for multiple clients, and in-house teams at mid-market B2B companies with defined ICP profiles.
Image generation quality depends heavily on how you prompt and iterate. We’ve put together a detailed guide on advanced prompting and reprompting techniques using two of our most popular image models: Flux 2 and Gemini 2.5 Flash Image (affectionately nicknamed ‘Banana’).
Flux 2 excels at artistic interpretation, stylistic consistency, and creative flexibility. It’s particularly strong when you want distinctive visual styles, illustration work, or conceptual imagery.
Gemini Banana leads in photorealism, coherence across remixing, and likeness preservation. It’s currently ranked number one on LMArena for both image editing and text-to-image generation, making it the go-to choice for product photography, marketing visuals, and when you need to maintain specific visual elements across iterations.
Starting Strong:
Iterative Refinement:
Advanced Controls:
Common Mistakes:
For Product Photography:
For Social Media Graphics:
For Campaign Hero Images:
We are working on a full guide and academy video tutorials which will include example prompts, before/after comparisons, and specific tips for common marketing use cases. Watch this space!
It’s that time of year. We’ve put together a short Christmas video showcasing what’s possible when you create entirely within Euryka – from concept to script to voiceover to video generation to final edit.
The video was made using:
No external tools. No context switching. Just one workspace with everything orchestrated around a single creative vision.
Look out for one more email before the year ends – our 2025 recap. But we’re still shipping updates, with plenty more happening in the background as we head into the new year.
Merry Christmas from all of us at Euryka. Thank you for being part of this journey.
We’re already working on the next set of updates, including:
As always, your feedback shapes our roadmap. If you have thoughts, questions, or want a walkthrough of any new features, reply to us via Intercom or book a quick session with the team.
All these features are available immediately to all subscribers:
If you’re on a trial, now’s the perfect time to upgrade and unlock the full power of what we’re building. Our Pro and Studio plans give you unlimited access to all models, templates, and features without restrictions.
