DeepSeek V4 Pro vs Kimi K3
Price, context and benchmark scores side by side — every number from the live model catalogue.
| DeepSeek V4 Prodeepseek | Kimi K3moonshotai | |
|---|---|---|
| Input / 1M tokens | $0.435 | $3.00 |
| Output / 1M tokens | $0.870 | $15.00 |
| 1M in + 300K out | $0.70 | $7.50 |
| Context window | 1.0M | 1.0M |
| Intelligence index | 44.3 | 57.1 |
| Coding index | 59.4 | 76.2 |
| Agentic index | 36.4 | 50.1 |
| Cheapest provider | DeepSeek | BaseTen |
| Providers serving it | 18 | 6 |
| Weights | Open | Open |
| Image input | No | Yes |
| Tool calling | Yes | Yes |
| Reasoning mode | Yes | Yes |
| Released | Apr 24, 2026 | Jul 16, 2026 |
Which one to pick
Pick DeepSeek V4 Pro if…
- Token cost matters: $0.870 per 1M output vs $15.00.
- You want provider redundancy: 18 providers serve it.
Pick Kimi K3 if…
- You want the higher general-intelligence score (57.1).
- Code quality is the priority — it leads on the coding index (76.2).
- You are running agent loops — higher agentic index (50.1).
- You need the bigger context window (1.0M).
FAQ
Is DeepSeek V4 Pro better than Kimi K3?
Kimi K3 scores higher on general intelligence (44.3 vs 57.1), while DeepSeek V4 Pro is 17× cheaper per output token ($0.870 vs $15.00). For code, Kimi K3 leads (76.2 coding index). Kimi K3 takes the larger context at 1.0M.
Which is cheaper, DeepSeek V4 Pro or Kimi K3?
DeepSeek V4 Pro, at $0.870 per 1M output tokens versus $15.00. On a 1M input + 300K output workload that is $0.70 against $7.50.
Can I self-host DeepSeek V4 Pro or Kimi K3?
DeepSeek V4 Pro: open weights (deepseek-ai/DeepSeek-V4-Pro). Kimi K3: open weights (moonshotai/Kimi-K3).