DeepSeek V4 Pro repriced across 2 hosts: Wafer input down 21%, cache read up 56%
Published 23 September 2026
Price move, per 1M tokens
- Wafer · input↓ 21%$0.500 → $0.395
- Wafer · outputunchanged$2.90 → $2.90
- Wafer · cache read↑ 56%$0.16 → $0.25
- NextBit · input↓ 6%$1.122 → $1.056
- NextBit · output↓ 6%$3.37 → $3.17
- NextBit · cache read↓ 5%$0.037 → $0.035
| host | rate | was | now | change |
|---|---|---|---|---|
| Wafer | input | $0.500 | $0.395 | ↓ 21% |
| Wafer | output | $2.90 | $2.90 | — |
| Wafer | cache read | $0.16 | $0.25 | ↑ 56% |
| NextBit | input | $1.122 | $1.056 | ↓ 6% |
| NextBit | output | $3.37 | $3.17 | ↓ 6% |
| NextBit | cache read | $0.037 | $0.035 | ↓ 5% |
Wafer input: ↓ 21%Wafer cache read: ↑ 56%NextBit input and output: ↓ 6%
If you buy
On Wafer, prompt-heavy work now costs less while repeated context costs more, so re-check your cache-hit share before your next top-up.
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 23 September 2026.