DeepSeek V4 Flash repriced across 2 hosts: Relace input and cache read down 50%, output up 300%
Published 30 September 2026
Price move, per 1M tokens
- Relace · input↓ 50%$0.018 → $0.009
- Relace · output↑ 300%$0.32 → $1.28
- Relace · cache read↓ 50%$0.018 → $0.009
- Inceptron · input↓ 11%$0.056 → $0.050
- Inceptron · outputunchanged$0.65 → $0.65
- Inceptron · cache readunchanged$0.027 → $0.027
| host | rate | was | now | change |
|---|---|---|---|---|
| Relace | input | $0.018 | $0.009 | ↓ 50% |
| Relace | output | $0.32 | $1.28 | ↑ 300% |
| Relace | cache read | $0.018 | $0.009 | ↓ 50% |
| Inceptron | input | $0.056 | $0.050 | ↓ 11% |
| Inceptron | output | $0.65 | $0.65 | — |
| Inceptron | cache read | $0.027 | $0.027 | — |
Relace input and cache read: ↓ 50%Relace output: ↑ 300%Inceptron input: ↓ 11%
If you buy
Relace now suits chatty, cache-heavy workloads far better than long generations, so re-check your output volume before your next top-up.
The model
DeepSeek V4 Flashopen the model →
Open weights290.9B parameters1049K context33 hosts
intelligence65th of 168 writing57th of 168 coding64th of 168 agents20th of 55
DeepSeek V4 Flash is a downloadable text model with a permissive licence, built for long-document work where the bill matters more than peak quality. Its measured quality is weak on the LiveBench boards and mid-pack on the Arena boards, so treat it as a cheap workhorse rather than a quality leader.
Source
Read on openrouter.ai · we published this on 30 September 2026.
read the original at openrouter.ai ↗