DeepSeek V3.1 by deepseek-ai — 684.53B MoE (37B active) open-weight local LLM. License: mit. Benchmarks, VRAM requirements per quantization, GPU compatibility, pricing and community reviews on slopsome.com.
Key specs
| Type | Local open-weight |
|---|
| Parameters | 684.53B total · MoE, 37B active |
|---|
| Architecture | deepseek_v3 |
|---|
| Context window | 164K tokens |
|---|
| Knowledge cutoff | 2025-06-01 |
|---|
| Modalities | text |
|---|
| Recommended backends | — |
|---|
| Minimum viable rig | — |
|---|
Benchmark scores
| GPQA Diamond | 74% |
|---|
| SWE-bench Verified | 66% |
|---|
| AIME | 88% |
|---|
| MMLU-Pro | 83% |
|---|
| BFCL v3 (tool use) | 75% |
|---|
| Composite score | 7.08 |
|---|
| Community rating | 4.5★ (2 reviews, 0 net votes) |
|---|
VRAM & disk per quantization
| Quant | VRAM | Disk | RAM | Context |
|---|
| Q8 | 700 GB | 690 GB | 768 GB | 131K |
| Q4_K_M | 398.5 GB | 397 GB | 448 GB | 164K |
API pricing (per 1M tokens)
| Provider | Input | Output | Free tier |
|---|
| DeepSeek | $0.27 | $1.1 | No |
| SiliconFlow | $0.27 | $1.1 | No |
| Together AI | $0.6 | $1.7 | No |
| Fireworks AI | $0.9 | $0.9 | No |