// DeepSeek · Text LLM
DeepSeek-V4-Flash-0731: VRAM & local deployment
304.2B (MoE) parameters. About 203 GB at 4-bit and 670 GB at BF16.
released 2026-07-31licence: Commercial OKmemory figures are estimates
01Specs
| Publisher | DeepSeek · deepseek-ai |
|---|---|
| Family / series | DeepSeek · DeepSeek-Flash |
| Type | Text LLM (text-generation) |
| Parameters | 304.2B (MoE) Source: counted from official safetensors weights |
| Context length | 1M tokens (1,048,576)max_position_embeddings in config.json |
| Architecture | 43 layers · GQA · MoE |
| Released weights | packed quantised tensors (I8, e.g. MXFP4 / INT4) · 167 GBdownload |
| Licence | MIT · Commercial OKPermissive licence (Apache-2.0 / MIT / CC-BY / OpenMDW): commercial use allowed. |
| Released | 2026-07-31 (Hugging Face repo created) |
02How much VRAM it needs (estimate)
Total = weights + 10% + 0.5GB overhead + KV cache for one request (FP16 KV). More concurrent users multiply the KV column.
| Precision | Weights | Smallest GPU | Mac |
|---|---|---|---|
| BF16 / FP16 (original) | 608 GB | 8 × H200 (1,128GB) | — |
| FP8 / INT8 / Q8_0 | 323 GB | 8 × H100 (640GB) | — |
| 4-bit (Q4_K_M / AWQ / GPTQ-Int4) | 184 GB | 4 × RTX PRO 6000 (384GB) | Mac 512GB |
KV cache not estimated: this model uses a new compressed / sparse attention design that our formula does not cover. Totals above are weights + overhead only.
Bits per weight: BF16 16, FP8 8.5, 4-bit 4.85. “Fits” = total ≤ 90% of memory; Mac GPU can use ~2/3 of unified memory (≤36GB) or ~3/4 (larger). For MoE, all experts must be in memory, so size by total parameters. Method →
03How to run it
Only runtimes the publisher or the runtime’s official library names for this model.
- vLLM — mentioned in the official model card
- SGLang — mentioned in the official model card
- Transformers — mentioned in the official model card
04Official repositories & quantised builds
1 official repo(s) from deepseek-ai.
05Other DeepSeek models
All 34 →| Model | Params | 4-bit total | Released |
|---|---|---|---|
| DeepSeek-V4-Flash · same series | 284B / 13B | 190 GB | 2026-04-22 |
| DeepSeek-V4.1-Flash-1017 · same series | 748B / 16B | 499 GB | 2026-10-17 |
| DeepSeek-V4.1-Flash · same series | 748B / 16B | 499 GB | 2026-09-10 |
| DeepSeek-V4-Pro-0813 | 1.65T | 1.10 TB | 2026-08-13 |
| DeepSeek-V4-Pro | 1.6T / 49B | 1.07 TB | 2026-04-22 |
| DeepSeek-OCR-2 | 3.39B | 3.3 GB | 2026-01-27 |
| DeepSeek-V3.2 | 685.4B | 458 GB | 2025-12-01 |
| DeepSeek-V3.2-Speciale | 685.4B | 458 GB | 2025-11-28 |
| DeepSeek-OCR | 3.34B | 3.2 GB | 2025-10-17 |
| DeepSeek-V3.2-Exp-Base | 685.4B | 458 GB | 2025-09-29 |
| DeepSeek-V3.1-Base | 671B / 37B | 449 GB | 2025-08-19 |
06Sources
Data checked 2026-08-01. Parameters, context and licence are copied from the official repository; memory figures are our estimates.