Updated daily at 05:00 UTC
Reasoning-only
DeepSeek-R
Reasoning models trained with large-scale reinforcement learning, starting with DeepSeek-R1.
By DeepSeek.
Versions
3 versions
| Model | Maker | Kind | Released | Context | Input $/1M | Output $/1M |
|---|---|---|---|---|---|---|
| DeepSeek-R1DeepSeek-R · Legacy · Open weights | DeepSeek | Reasoning-only | 20 Jan 2025 | 131.1K | — | — |
| DeepSeek-R1-ZeroDeepSeek-R · Legacy · Open weights | DeepSeek | Reasoning-only | 20 Jan 2025 | 131.1K | — | — |
| DeepSeek-R1-Lite-PreviewDeepSeek-R · Retired | DeepSeek | Reasoning-only | 20 Nov 2024 | — | — | — |
