$ diff deepseek-v4-flash kimi-k3
DeepSeek V4-Flash vs Kimi K3
Values come from DeepSeek V4-Flash and Kimi K3, where each is cited to its source. This page states no benchmark result and ranks neither model — it puts two published specifications next to each other. Where a lab has not published a figure, the row says so rather than guessing.
DeepSeekDeepSeek V4-Flash
- context
- —
- weights
- open
- $/M in
- $0.14
- $/M out
- $0.28
- context
- 1M
- weights
- open
- $/M in
- $3
- $/M out
- $15
What actually differs
- Input price
- DeepSeek V4-Flash at $0.14/M against $3/M — 21.4× cheaper to feed. Standard rates: one of these labs also quotes a lower cached-input tier, which applies only when a prefix is reused — see the full spec below.
- Output price
- DeepSeek V4-Flash at $0.28/M against $15/M — 53.6× cheaper to generate. Output dominates the bill on most agentic workloads, where the model writes far more than it reads.
- Weights
- Both publish weights — DeepSeek V4-Flash under MIT (open-weight), Kimi K3 under Modified MIT (commercial use permitted; weights publicly downloadable). Check the licences rather than assuming they permit the same commercial use.
- Recency
- DeepSeek V4-Flash shipped 15 days after Kimi K3 (2026-07-31 vs 2026-07-16).
Full spec
| Attribute | DeepSeek V4-Flash | Kimi K3 |
|---|---|---|
| Developer | DeepSeek | Moonshot AI |
| Released | 2026-07-31 | 2026-07-16 |
| Context window | not recorded | 1M tokens |
| Pricing | $0.14/M input (cache miss) · $0.0028/M input (cache hit) · $0.28/M output — see Conflicting Reports | $0.30/M cache-hit input · $3.00/M cache-miss input · $15.00/M output |
| License | MIT (open-weight) | Modified MIT (commercial use permitted; weights publicly downloadable) |
| Availability | DeepSeek API (Responses format, Codex-adapted), Hugging Face (`deepseek-ai/DeepSeek-V4-Flash-0731`) | Kimi Code, Kimi app, Kimi API; weights on Hugging Face (`moonshot-ai/kimi-k3`, released 2026-07-27) |