DeepSeek-V4-Flash-0731-NVFP4
> [!IMPORTANT] > This is a community conversion of the official DeepSeek V4 Flash 0731 > checkpoint. DeepSeek's architecture, tokenizer, DSpark/MTP configuration, > attention, shared experts, router, head, and other excluded tensors are > preserved. This repository is not affiliated with or endorsed by DeepSeek or > NVIDIA.
Params
304.2 B
Context
1,048,576
Downloads 30d
203 K
Likes
18
Download history
daily snapshots · 38 days
▲ 104 K in the last 30 days (33.8%)
458 K203 K
Aug 22Sep 1Sep 11Sep 20
458 K175 K
Aug 14Aug 26Sep 8Sep 20
Can you run it?
Estimated VRAM at 8K context unless noted. Pick your hardware to see the verdict per quantization.
| File | Quant | Size | Est. VRAM | Verdict on RTX 4090 · 24 GB |
|---|---|---|---|---|
| model.safetensors | u8 | 175.6 GB | 239.2 GB | ❌ Won’t fit |
| model.safetensors (bf16, full) | bf16 + 1024K ctx | 175.6 GB | 6,033.9 GB | ❌ Won’t fit |
Estimate: file size × 1.1 + KV cache at 8K + 0.5 GB overhead. Not a benchmark — how we calculate this.
Run it
copy-paste, exact tags checked against the Hub$ curl -s https://aimodelscomparison.com/api/v1/models/deepseek-v4-flash-0731-nvfp4
{
"hf_id": "MJPansa/DeepSeek-V4-Flash-0731-NVFP4",
"params_b": 304.18,
"context_length": 1048576,
"license": { "id": "mit", "commercial": "yes" },
"downloads_30d": 202545,
"vram_estimates": [
{ "quant": "u8", "gb": 239.2 }
],
"updated_at": "2026-08-14T01:00:47Z"
}
Specifications
- Architecture
- DeepseekV4ForCausalLM
- Parameters
- 304.2 B
- Tensor type
- U8
- Context length
- 1,048,576
- Vocabulary
- 129,280
- Layers / heads
- 43 / 64
- Licence
- mit
- First seen on the Hub
- 2026-07-31
- Base model
- DeepSeek-V4-Flash-0731
- Training datasets
- undisclosed
- Added to our catalog
- 2026-08-14
Family
Base model and the most-downloaded derivatives in the catalog.
Compare with any text-generation model