DeepSeek V4.1 Flash repriced across 3 hosts: DekaLLM output down 33%, Relace output up 50%
Published 27 September 2026
Price move, per 1M tokens
- Relace · inputunchanged$0.050 → $0.050
- Relace · output↑ 50%$0.40 → $0.60
- Relace · cache read↑ 100%$0.005 → $0.010
- DekaLLM · input↓ 20%$0.15 → $0.12
- DekaLLM · output↓ 33%$0.60 → $0.40
- DekaLLM · cache readunchanged$0.005 → $0.005
- Wafer · input↓ 14%$0.049 → $0.042
- Wafer · outputunchanged$0.60 → $0.60
- Wafer · cache read↓ 16%$0.045 → $0.038
| host | rate | was | now | change |
|---|---|---|---|---|
| Relace | input | $0.050 | $0.050 | — |
| Relace | output | $0.40 | $0.60 | ↑ 50% |
| Relace | cache read | $0.005 | $0.010 | ↑ 100% |
| DekaLLM | input | $0.15 | $0.12 | ↓ 20% |
| DekaLLM | output | $0.60 | $0.40 | ↓ 33% |
| DekaLLM | cache read | $0.005 | $0.005 | — |
| Wafer | input | $0.049 | $0.042 | ↓ 14% |
| Wafer | output | $0.60 | $0.60 | — |
| Wafer | cache read | $0.045 | $0.038 | ↓ 16% |
Relace output: ↑ 50%DekaLLM output: ↓ 33%Wafer input: ↓ 14%
If you buy
Relace now costs more to run for output-heavy work, so check whether your usage leans on output or cache reads before your next top-up there.
The model
DeepSeek V4.1 Flashopen the model →
Open weights763.2B parameters1049K context35 hosts
intelligence21st of 168 writing37th of 168 coding11th of 168 agents13th of 55
DeepSeek V4.1 Flash is a downloadable model with a permissive licence, and it places 11th of 168 on Arena Coding via Max as of 25 Sep 2026. At 763.2 billion parameters it is a data-centre job to run yourself, so hosted use is the practical route for most readers.
Source
Read on openrouter.ai · we published this on 27 September 2026.
read the original at openrouter.ai ↗