models · catalogue
8 open source models, served in Europe.
Open models deployed on NVIDIA B200 GPUs in Spain, with an API compatible with the OpenAI and Anthropic SDKs. From 0.03 € per million tokens, with your data in the EU.
DeepSeek V4 Pro★
DeepSeek AI
1.6T MoE on 8× B200. Full-power reasoning.
1.6T MoE8× B20032,768 contextFP4View modelinput / output · €/M
0.42 / 0.85 €
GLM 5.2 FP8★
Z.ai
754B FP8 on 8× B200. Generalist and coding with tool calling.
743B · 39B active MoE8× B200202,752 contextFP8View modelinput / output · €/M
1.00 / 3.00 €
DeepSeek V4 Flash
DeepSeek AI
1.6T MoE on 2× B200. High throughput, 524K context.
1.6T MoE2× B200524,288 contextFP8View modelinput / output · €/M
0.11 / 0.22 €
Qwen3.6 27B
Alibaba Cloud
27B dense. Strong at coding and agents, with tool calling.
27B denseMIG B20098,304 contextFP8View modelinput / output · €/M
0.10 / 0.30 €
Qwen3 14B
Alibaba Cloud
14B FP8. Reasoning with a configurable thinking budget.
14BMIG B20032,768 contextFP8View modelinput / output · €/M
0.10 / 0.24 €
Qwen3.5 9B
Alibaba Cloud
9B ultra-light. Fast and cost-effective, with tool calling.
9BMIG B20065,536 contextFP8View modelinput / output · €/M
0.03 / 0.15 €
Gemma 4 E4B
Google DeepMind
MoE, 4B active params. Lightweight, reasoning and tools.
4B active · MoEMIG B200131,072 contextBF16View modelinput / output · €/M
0.10 / 0.28 €
Gemma 4 E2B
Google DeepMind
MoE, 2B active params. Smallest and most affordable.
2B active · MoEMIG B200131,072 contextBF16View modelinput / output · €/M
0.05 / 0.24 €
NVIDIA B200 GPUs · Spain (EU)
Ready to call them?
Create your account with 5 € free, point your SDK at GPU Flow and start in minutes.