EbbotGPT LLMs
Learn about the available LLMs in EbbotGPT and use them for to build AI agents for your service operations.
On the page you can read about the technical details and practical implications of our in-house models in EbbotGPT. For more technical documentation about the additional LLMs that we support in EbbotGPT, we refer to the AI providers docs.
These LLMs can be used when building agents regardless if you're building a chat agent, an email agent or using the EbbotGPT API.
Available in-house models in EbbotGPT
We regularly release new and improved LLM versions. Our currently available models are:
EbbotGPT 3 - Released 3/11-2025
EbbotGPT 3
Knowledge comprehension
EbbotGPT 3 has almost twice the parameters of EbbotGPT 2 which improves its knowledge comprehension and problem-solving capabilities.
Speed
Despite being bigger than EbbotGPT 2, EbbotGPT 3 is twice as fast because it uses a Mixture of Experts (MoE) architecture. This means the model only activates the most relevant 5% of its parameters to answer a user's query.
Integrated tool calling
EbbotGPT 3 features integrated tool calling, enabling it to seamlessly decide whether to use a tool or generate an answer from the uploaded knowledge.
Context window
With a context window of 31,000 tokens, EbbotGPT 3 can process and manage four times more information than EbbotGPT 2.
2
6
7
3
8
9
Additional models available in EbbotGPT
It is possible to use external LLMs in EbbotGPT. See what additional, external models that are available below.
OpenAI GPT (Azure)
Google AI
For more technical information on model performance, follow the link below.
Model ComparisonsChoosing between EbbotGPT LLMs and API LLMs
Ebbot’s Sovereign LLMs
Ebbot’s models, built on powerful open-source foundations are the primary choice when data residency and legal compliance are non-negotiable. Because these models are hosted entirely within the EU, they are designed to align with strict GDPR and EU AI Act standards. Rather than being generalists, these models are tested and prompt-optimized specifically for service automation and high-stakes customer interactions. You should choose these when you need a "domain expert" that is deeply integrated with your corporate knowledge through Retrieval-Augmented Generation (RAG), as they offer superior control over model behavior and significantly lower the risk of irrelevant responses in a business context.
Global API LLMs
Conversely, you should opt for Global API LLMs when your project requires massive scale, native multimodality, or an exceptionally large context window. These cutting-edge commercial models excel at broad, complex reasoning, such as analyzing massive amounts of documentation in a single pass. They are ideal for high-volume research, creative content generation, or experimental prototyping where the absolute latest "frontier" capabilities are more critical than localized hosting or specific service-industry optimization.
The Quick Rule of Thumb:
Use Ebbot’s models for sensitive, customer-facing service workflows that require EU data sovereignty and verified, service-ready accuracy.
Use Global APIs for broad-scale data processing, complex multi-step reasoning, or if you have any specific use-case that EbbotGPT LLMs struggle with.
EbbotGPT LLM
Global API LLMs
Data Location
🇪🇺 Strictly EU-based
🌍 Global / US-based
Capacity
Medium
Very big (1M+ tokens)
Optimization
Prompt-optimized for service
General-purpose reasoning
Speed
Medium
High
Oversight
Full Control over hosting
No Control (Black-box updates)
Combine EbbotGPT LLMs with API LLMs
You can combine these LLMs if you want to optimize for both security and versatility. By routing your base use-case, such as daily customer support and sensitive internal data processing, EbbotGPT LLMs you maintain full data sovereignty and operational reliability. You can then call upon global APIs as a secondary layer for specialized, high-intensity reasoning tasks that falls outside the scope of routine service logic.
Last updated
Was this helpful?