DeepSeek-V4.1-Flash
We introduce DeepSeek-V4.1-Flash, a multimodal Mixture-of-Experts (MoE) model with 552B backbone parameters and support for contexts of up to one million tokens. The model natively processes images and text, and generates text autoregressively.
Params
763.2 B
Context
—
Downloads 30d
1.5 M
Likes
4,716
Download history
daily snapshots · 20 days1.5 M798 K
Oct 5Oct 11Oct 18Oct 24
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 | 510.3 GB | 676.3 GB | ❌ Won’t fit |
Estimate: file size × 1.1 + KV cache at 8K + 0.5 GB overhead. Not a benchmark — how we calculate this.
Use via API
via OpenRouter · USD per 1M tokens · synced 2026-10-24$0.02
input / 1M
$1.00
output / 1M
$0.015
cached input / 1M
$1.02
1M in + 1M out
:batch variant
$0.05 in · $0.15 out
- Context window
- 1,048,576 tokens
- Max output
- 943,718 tokens
- Providers
- 33
- Cheapest provider
- Baidu · $0.054 / $0.21
- Best throughput
- —
- AA Intelligence / Coding / Agentic
- 24.7 / — / —
✓ Tool calling
✓ Reasoning
✓ Structured outputs
✓ Image input
✗ File input
✗ Audio input
✓ Prompt caching
Specifications
- Architecture
- DeepseekV41ForCausalLM
- Parameters
- 763.2 B
- Tensor type
- I8
- Licence
- mit
- First seen on the Hub
- 2026-09-10
- Training datasets
- undisclosed
- Added to our catalog
- 2026-10-05
Compare with any image-text-to-text model
Popular comparisons
Appears in