DeepSeek
Hangzhou lab whose V and R series models are released under permissive licences and undercut Western API pricing by an order of magnitude.
Full DeepSeek record →Head to head
DeepSeek vs Kimi
DeepSeek (DeepSeek) and Kimi (Moonshot AI) both sit in Writing. Every value below comes from the latest crawl of the vendors' own pages.
| Attribute | DeepSeek DeepSeek | Kimi Moonshot AI |
|---|---|---|
| Segment | Writing | Writing |
| Version | DeepSeek-V4.1-Flash | Kimi K3 |
| Entry cost | Pay-as-you-go API pricing based on token consumption. No flat monthly fee. | Pay-as-you-go based on token usage. See pricing tables for Kimi K3, K2.7, and K2.6 models. |
| Pricing tiers | deepseek-flash (Input Cache Hit - Off-Peak) $0.003 · deepseek-flash (Input Cache Hit - Peak) $0.006 · deepseek-flash (Input Cache Miss - Off-Peak) $0.15 · deepseek-flash (Input Cache Miss - Peak) $0.3 · deepseek-flash (Output - Off-Peak) $0.6 · deepseek-flash (Output - Peak) $1.2 | Pay-as-you-go API usage free |
| Model stack | DeepSeek-R1 · DeepSeek-V3.2 · deepseek-v4-flash · DeepSeek-V4-Flash-0731 · DeepSeek-V4-Flash-Vision-Exp · DeepSeek-V4-Pro · DeepSeek-V4-Pro-0813 · DeepSeek-V4.1-Flash | Kimi K2.6 · Kimi K2.7 Code · Kimi K3 · kimi-k2.5 · kimi-k2.7-code-highspeed |
| Context window | 1M | 1M |
| Public API | ||
| Routes models | ||
| Changes / 30d | 9 | 9 |
| Origin | China | China |
| Capabilities | Reasoning mode, Open weights (MIT), OpenAI-compatible API, Context caching, Function calling, Thinking Mode, Context Caching, JSON Output | Native multimodal (visual + text), 1M-token context, Agentic tool use, Thinking / non-thinking modes, Dedicated coding models (k2.7-code, high-speed), Open weights, 1M token context, Multimodal input |
Kimi
Moonshot AI's assistant, known for very long context windows and the open-weight K2 agentic model.
Full Kimi record →