DeepSeek R1 Distill Llama 70B — Specifications
| Developer | DeepSeek |
|---|
| Type | LLM (dense, reasoning) |
|---|
| Modality | Text → Text |
|---|
| Parameters | 70B |
|---|
| Context window | 128K |
|---|
| Max output | — |
|---|
| License | MIT (open) |
|---|
| Open weights | ✅ Yes |
|---|
| Released | 2025 |
|---|
| Input price | $0.80 /1M |
|---|
| Output price | $0.80 /1M |
|---|
| API providers | DeepInfra, OpenRouter, Ollama |
|---|
🖥️ Run it locally
| VRAM (4-bit) | ~40 GB |
|---|
| Minimum GPU | 2× RTX 4090 / 1× 48GB |
|---|
Official page →
R1’s reasoning distilled into a 70B Llama base — brings much of DeepSeek R1’s step-by-step reasoning to hardware you can actually self-host. MIT-licensed, 128K context.
