DeepSeek-V4-Flash vs KAT-Coder-Pro V2
DeepSeek vs Kuaishou — benchmarks, pricing, and capabilities side by side.
- •DeepSeek-V4-Flash has the higher intelligence index (89.4 vs 85.5)
- •DeepSeek-V4-Flash is cheaper ($0.10 vs $0.30 per 1M input)
- •DeepSeek-V4-Flash is faster
- •DeepSeek-V4-Flash has a larger context window (1M)
| DeepSeek-V4-Flash | KAT-Coder-Pro V2 | |
|---|---|---|
| Intelligence index | 89.4 | 85.5 |
| Developer | DeepSeek | Kuaishou |
| Type | LLM | LLM |
| Access | Open weights | API only |
| Context window | 1,048,576 tokens | 256,000 tokens |
| Input price | $0.10 / 1M | $0.30 / 1M |
| Output price | $0.20 / 1M | $1.20 / 1M |
| Speed | 109 tok/s | 108 tok/s |
| Released | April 24, 2026 | March 27, 2026 |
| Parameters | 284B (13B active) | — |
| Input modalities | Text | Text |
| Output modalities | Text | Text |
Shared benchmarks
DeepSeek-V4-Flash
KAT-Coder-Pro V2
GPQA Diamond
89.4
85.5
Humanity’s Last Exam
32.1
16