Host AtlasCloud cut DeepSeek V3.2 input and cache-read pricing by 48%
Published 28 September 2026
Price move, per 1M tokens
- AtlasCloud · input↓ 48%$0.260 → $0.134
- AtlasCloud · output↓ 47%$0.38 → $0.20
- AtlasCloud · cache read↓ 48%$0.130 → $0.067
| host | rate | was | now | change |
|---|---|---|---|---|
| AtlasCloud | input | $0.260 | $0.134 | ↓ 48% |
| AtlasCloud | output | $0.38 | $0.20 | ↓ 47% |
| AtlasCloud | cache read | $0.130 | $0.067 | ↓ 48% |
input and cache read: ↓ 48%
The model
DeepSeek V3.2open the model →
Open weights685.4B parameters164K context17 hosts
intelligence74th of 168 writing68th of 168 coding75th of 168
DeepSeek V3.2 is a large downloadable text model built for coding and reasoning, with 685.4 billion total parameters and 37 billion active per word. It carries a permissive MIT licence and is available from 20 hosted providers.
Source
Read on openrouter.ai · we published this on 28 September 2026.
read the original at openrouter.ai ↗