Lyceum Technology

We cut GLM-5.3 costs by 20%

Our new inference engine is faster, so you pay less. Plus a new website and a new logo.

DeepSeek V4.1 Flash is live

DeepSeek V4.1 Flash is now available on Lyceum

Lyceum in PowerPoint, Claude Code preconfigured, and new B200/B300 capacity

Lyceum in your slides, your terminal, and more GPU capacity in Europe

33 models, prompt caching, and more H100s

The model catalog has grown to 33 models (incl. DeepSeek V4 Pro, GLM-5.2, and Kimi K2.7 Code). Prompt caching cuts input costs by up to 83% and more H100 nodes are available

Serverless inference is live. The on-demand GPUs are back

Pay-per-token access to open-source models on EU-sovereign GPUs. Plus: on-demand H100, H200, B200 & B300 capacity is back, and reserved nodes are on the way.

The inference gap in European AI - and how we're closing it