Tuned serving for open-weight models — GLM, Kimi and DeepSeek.
Wafer serves open-weight models on a stack it tunes per workload, profiling traffic to pick the serving engine, kernels and hardware for each deployment. US-hosted. The route to pick when you want tuned throughput on GLM, Kimi or DeepSeek and US hosting is acceptable.
Every model we route through Wafer. Compare residency, zero retention, training posture, and price at a glance — full data-handling detail per route below.
| Model | Region | Zero data retention | Training | Context | Input | Output |
|---|---|---|---|---|---|---|
| US | Zero data retention | No | 1M | $3.00 | $12.75 | |
| US | Zero data retention | No | 1M | $1.19 | $4.40 | |
| US | Zero data retention | No | 1M | $0.10 | $0.35 | |
| US | Zero data retention | No | 1M | $1.26 | $3.96 | |
| US | Zero data retention | No | 1M | $0.10 | $0.25 |
Wafer hosts on 1 route. Each route has its own privacy posture, residency, and no-KYC terms. Postures are maintained by qynio with a last-verification timestamp.
Zero data retention is on by default on Pay-as-you-go — no action required. No training on customer data. US; wallets; deposit available.
