Skip to content
- New host promotion — machines registered in 2026 lock in our lowest 7.5% fee for 6 months. Fee tiers →
- Reliability tier thresholds were recalibrated, and your machine’s current tier and fee are shown directly in the host console.
- Pick your context length at deploy time. LLM deployments let you choose the max context length interactively, with GPU recommendations updating live — shorter contexts unlock cheaper GPUs.
- VRAM requirements are estimated more accurately for long-context deployments, so a deployment that fits actually fits.
- Viewing a workspace that was already terminated now shows a clear “not found” instead of an error.
