DeepSeek V4-Flash
DeepSeekTEXT Open weight
- Provider
- DeepSeek
- Modality
- TEXT
- Parameters
- 284B
- Context window
- 1,000,000 tokens
- Weights
- Open
- Released
- 24 Apr 2026
DeepSeek V4-Flash is the cheap, fast 1M-context MoE — extraordinary tokens-per-dollar that makes it the budget default for high-volume Indian deployments.
Best for
- Highest tokens-per-dollar; default for chat / RAG / high volume
- 1M context at roughly one-third the V4-Pro price
- Near-free cache-hit input for repeated prompt prefixes
How it compares — and the India angle
- The value pick for Indian startups: ~$0.14 in / $0.28 out is roughly 10-30x below comparable Western frontier APIs.
- Structure prompts with stable system-prompt prefixes to ride the ~$0.003/1M cache-hit rate.
- MIT-licensed and far smaller (284B) than V4-Pro — a realistic self-host / data-residency option.
How to access
API ~$0.14 / 1M cache-miss input (~$0.003 cache-hit), ~$0.28 / 1M output. MIT (free to self-host).
Access DeepSeek V4-Flash