DeepSeek
Open-weight frontier models with a focus on reasoning and cost efficiency
DeepSeek models are the default choice for high-volume workloads: strong reasoning, long context and a price point that makes them viable as an always-on backbone. On StarSeaAPI every DeepSeek variant is served through multiple independent upstream routes, so a single provider incident does not take the model offline.
Models (6)
| Models | Input $/M | Output $/M | Intelligence index | Coding index |
|---|---|---|---|---|
| deepseek-flash | $0.15 | $0.60 | — | — |
| deepseek-v4-flash | $0.15 | $0.60 | 34.3 | 69.1 |
| deepseek-v4-flash-0731 | $0.15 | $0.60 | 34.3 | 69.1 |
| deepseek-v4.1-flash | $0.15 | $0.60 | 39.5 | — |
| deepseek-v4-pro | $0.66 | $1.98 | 36.0 | 68.8 |
| deepseek-v4-pro-0813 | $0.66 | $1.98 | 36.0 | 68.8 |
Prices are per 1M tokens in USD · ¥6.5 = $1