DeepSeek-V4-Pro-0813
DeepSeek-V4-Pro-0813 is the official release of DeepSeek-V4-Pro, superseding the preview version, with greatly enhanced agentic capabilities and performance improvements that are especially pronounced in production environments. It is built on the DeepSeek-V4-Pro (Preview) model structure, with a DSpark speculative decoding module attached.
Params
1,650.5 B
Context
1,048,576
Downloads 30d
150 K
Likes
899
Download history
tracking started — chart appears after 7 days of snapshots (5 recorded)150 K downloads in the last 30 days
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 | i8 | 892.7 GB | 1,230.1 GB | ❌ Won’t fit |
| model.safetensors (bf16, full) | bf16 + 1024K ctx | 892.7 GB | 32,672.1 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-pro-0813
{
"hf_id": "deepseek-ai/DeepSeek-V4-Pro-0813",
"params_b": 1650.50,
"context_length": 1048576,
"license": { "id": "mit", "commercial": "yes" },
"downloads_30d": 150373,
"vram_estimates": [
{ "quant": "i8", "gb": 1230.1 }
],
"updated_at": "2026-10-19T06:02:33Z"
}
Use via API
via OpenRouter · USD per 1M tokens · synced 2026-10-24$0.66
input / 1M
$1.98
output / 1M
$0.022
cached input / 1M
$2.64
1M in + 1M out
Price tiers
- Weekdays 01:00–04:00 and 06:00–10:00 UTC: input $1.32/M, output $3.96/M (2× base)
- Context window
- 1,048,576 tokens
- Max output
- 393,216 tokens
- Providers
- 21
- Cheapest provider
- Baidu · $0.13 / $0.40
- Best throughput
- —
- AA Intelligence / Coding / Agentic
- 36 / 68.8 / 41.3
✓ Tool calling
✓ Reasoning
✓ Structured outputs
✗ Image input
✗ File input
✗ Audio input
✓ Prompt caching
Specifications
- Architecture
- DeepseekV4ForCausalLM
- Parameters
- 1,650.5 B
- Tensor type
- I8
- Context length
- 1,048,576
- Vocabulary
- 129,280
- Layers / heads
- 61 / 128
- Licence
- mit
- First seen on the Hub
- 2026-08-13
- Training datasets
- undisclosed
- Added to our catalog
- 2026-10-20
Compare with any text-generation model