Model names are perishable facts. What is at the top today drops a tier within weeks, and comparisons that held six months ago now mislead. So from here on we track the changes to the novendix.AI model line-up in this place – with a date, so it stays clear what any statement refers to.
This overview is current as of September 2026.
What changed most recently
New top models in the Ultra tier
The Ultra tier now carries Claude Opus 5.5 with a context window of one million characters – the highest quality on complex questions. Also in Ultra: GPT 5.6 Sol for structured data. Claude Fable 5.1 was briefly planned but is not enabled: the model does not meet our binding requirement that providers store no requests, so our gateway rejects it.
Claude Sonnet 5 in the Premium tier
Premium now carries Claude Sonnet 5 as the balanced everyday model; the previous version has been dropped. Gemini 3.8 Flash joins it as a fast model of the latest generation.
Five additional models
Newly added are Gemini 3.5 Flash Lite, Gemini 2.5 Flash Lite, GPT-5 Nano, Gemini 2.5 Pro and Mistral Medium 3.5 – the only Mistral model with image recognition and reasoning. The three Lite and Nano variants are aimed above all at simple, frequent tasks where speed and cost matter.
The line-up at a glance
That makes nineteen language models across three performance tiers: Standard with ten, Premium with seven and Ultra with two, plus two automatic selectors for text and images. The full list with provider and context size is on AI models at a glance; which model family proves itself for what in practice is set side by side in the LLM comparison.
Why a model appears here later
A new model is not enabled the moment it is announced, but only once it is contractually guaranteed that requests are neither used for training nor stored after the answer has been given, and once processing can take place in a German data centre. Processing on our behalf under Art. 28 GDPR has to be settled as well.
That is why our list is shorter than some providers’ – and why weeks sometimes pass between a model being announced and it becoming available here. What that means in detail is on Privacy & Security.
What this section covers
Three kinds of item, each kept short and written with one question in mind: what does it practically mean for a mid-sized company?
- changes to the model line-up – models added, models dropped, tiers moved
- new platform features and what they replace in day-to-day work
- legal changes affecting the use of AI, for instance from the GDPR and the AI Act
No announcements without substance and no benchmark tables: if a change makes no difference to daily work, we say so.
Want to know if this pays off in your company? We’ll look at a concrete process with you and tell you honestly even if it isn’t worth it.
Your secure AI platform for the Mittelstand. Secure. Intelligent. Integrated. Custom database integration, personally supported.
novendix GmbH · Industriestraße 6 · 91126 Schwabach
Locations: Schwabach · Weißenburg · Nuremberg
A company of the L&S Lange & Schermer Group
