Models in catalogue
456
406 EU-runnable · 2 warm right now
Running GPU instances
3
across all customers and pool instances
Scale-to-zero
On
Cold models auto-boot · 15m idle timeout
Currently warm
| Model | Origin | Min VRAM | Token price (€/M in / out) |
|---|---|---|---|
Qwen2.5 14B Instruct AWQ Qwen/Qwen2.5-14B-Instruct-AWQ |
Qwen | 34 GB | 0.1500 / 0.2500 |
Qwen3 8B Qwen/Qwen3-8B |
Qwen | 19 GB | 0.1000 / 0.1800 |
Incidents (last 30 days)
No incidents to report.
Roadmap status
- EU Hosted Gateway (Vast.ai EU): live beta
- Single-Tenant Encrypted tier (at-rest + BYOK): in development
- Scaleway sovereign GPUs: integration starting
- Confidential computing (TEE / encrypted VRAM): roadmap — gated on EU provider attestation
- ISO 27001 certification: trajectory started
Snapshot taken 2026-06-25T18:27:19+00:00 (1 seconde geleden).