DeepSeek-V3.1 — reviews, specs & pricing
DeepSeek's hybrid reasoning MoE model.
Summary
DeepSeek-V3.1 is a 671B parameter (37B active) mixture-of-experts model combining fast and thinking modes in one model. It offers strong coding and reasoning at very low API cost. It's released with open weights for research and commercial use.
Sample use case
Used for cost-sensitive high-volume inference, coding agents, and research on MoE architectures. Popular for self-hosting on large GPU clusters.
Specifications
- Provider: deepseek
- License: open
- Parameters: 671B
- Context: 128k tokens
- Input price: $0.56/M tok
- Output price: $1.68/M tok
- Released: 2025-08-21
Pros
- Very low cost
- Open weights
- Hybrid think/non-think modes
Cons
- Requires large compute to self-host
- Less polished tooling ecosystem
- Data/privacy concerns for hosted API
Average rating 0.0 from 0 community reviews on Reviuws.