DeepSeek-R1
by DeepSeek
· DeepSeek R1 family · released Jan 20, 2025
Open weights
Active
Open reasoning matching o1, trained by pure RL.
Pricing
Launch price (from Jan 20, 2025)
$0.550 in / $2.19 out per 1M tokens
Current price (since Jan 20, 2025)
$0.550 in / $2.19 out per 1M tokens
Cache read: $0.140/M
| Tier | Effective | In $/M | Out $/M | Cache $/M | Per image | Per second | Source |
|---|---|---|---|---|---|---|---|
| Standard | Jan 20, 2025 – now | $0.550 | $2.19 | $0.140 | - | - | - |
| Batch | Jan 20, 2025 – now | $0.280 | $1.10 | - | - | - | - |
Specifications
| Model type | Text LLM |
|---|---|
| Architecture | MoE |
| Parameters | 671 (37 active) |
| Context window | 131.1K tokens |
| Max output | 8.2K tokens |
| Knowledge cutoff | Aug 1, 2024 |
| License | MIT License |
| Modalities | Text input Text output |
| Tags | Flagship Reasoning |
| End of life | Not disclosed |
| Last verified | Aug 20, 2026 |
Benchmarks
| Benchmark | Score | Src |
|---|---|---|
| LMSYS Chatbot Arena Elo | 1,358 | - |
| MATH | 90.8% | official |
| GPQA Diamond | 71.5% | official |