Qwen3.8 27B repriced across 2 hosts: Wafer input down 17%, output up 76%
Published 23 September 2026
Price move, per 1M tokens
- Wafer · input↓ 17%$0.120 → $0.099
- Wafer · output↑ 76%$2.50 → $4.40
- Wafer · cache read↑ 78%$0.050 → $0.089
- DekaLLM · input↓ 17%$0.12 → $0.10
- DekaLLM · outputunchanged$2.50 → $2.50
- DekaLLM · cache readunchanged$0.050 → $0.050
| host | rate | was | now | change |
|---|---|---|---|---|
| Wafer | input | $0.120 | $0.099 | ↓ 17% |
| Wafer | output | $2.50 | $4.40 | ↑ 76% |
| Wafer | cache read | $0.050 | $0.089 | ↑ 78% |
| DekaLLM | input | $0.12 | $0.10 | ↓ 17% |
| DekaLLM | output | $2.50 | $2.50 | — |
| DekaLLM | cache read | $0.050 | $0.050 | — |
Wafer input: ↓ 17%Wafer output: ↑ 76%DekaLLM input: ↓ 17%
If you buy
Output-heavy work at DekaLLM now costs more to run, so check your token mix before your next top-up and shift long generations elsewhere if input dominates your bill.
The model
Qwen3.8 27Bopen the model →
Open weights27.8B parameters262K context20 hosts
intelligence63rd of 168 writing93rd of 168 coding45th of 168 agents33rd of 55
Qwen3.8 27B is a downloadable model you can run yourself, and its measured strength is maths and reasoning. Agentic recovery and tool use sit near the bottom of the field, so it suits a solver rather than an agent that has to get itself back on track.
Source
Read on openrouter.ai · we published this on 23 September 2026.
read the original at openrouter.ai ↗