Host Wafer repriced DeepSeek V4 Pro: input down 43%, cache read up 300%
Published 21 September 2026
Price move, per 1M tokens
- Wafer · input↓ 43%$0.98 → $0.56
- Wafer · outputunchanged$2.90 → $2.90
- Wafer · cache read↑ 300%$0.033 → $0.132
| host | rate | was | now | change |
|---|---|---|---|---|
| Wafer | input | $0.98 | $0.56 | ↓ 43% |
| Wafer | output | $2.90 | $2.90 | — |
| Wafer | cache read | $0.033 | $0.132 | ↑ 300% |
input: ↓ 43%cache read: ↑ 300%
If you buy
If you lean on cached context, re-check your typical mix before your next top-up, since prompt-heavy work now costs less while repeated reads cost more.
The model
DeepSeek V4 Proopen the model →
Open weights1598.8B parameters1049K context28 hosts
intelligence36th of 168 writing32nd of 168 coding43rd of 168 agents31st of 55
DeepSeek V4 Pro is a large downloadable text model with a one-million-token request limit and strong mathematics scores. Its mixture-of-experts design keeps only 37 billion parameters active per token, making it more efficient to run than its total size suggests.
Source
Read from the hosts' own published rates · we published this on 21 September 2026.