LLM
GA
Open source
DeepSeek V4 Flash
Cheap, fast and capable open model; high volume
Release2026-08
Context1M
Max output65,536 tokens
Parameters284B
Local / VRAM280–400 GB
LicenseOpen source
Overview
DeepSeek V4 Flash is a model in the LLM category from DeepSeek. It was released in 2026-08. The vendor is based in China. Best for: Cheap, fast and capable open model; high volume.
Strengths
- 284B/13B
- 1M context
- MIT
- Very cheap
Limits
- Reasoning not as deep as Pro
- No vision
Specifications
| Vendor | DeepSeek |
|---|---|
| Category | LLM |
| Release | 2026-08 |
| Status | GA |
| License | Open source |
| License name | MIT |
| Commercial use | Allowed |
| Parameters | 284B |
| Active parameters | 13B |
| Architecture | MoE transformer |
| Context window | 1M |
| Max output tokens | 65,536 |
| Input modalities | Text |
| Output modalities | Text |
| Prompt types | Text |
| Negative prompts | No |
| Max outputs | 1 |
| Local deployment | Yes |
| VRAM | 280–400 GB |
| System RAM | 320–640 GB |
| API access | Yes |
| Pricing | $0.14 / $0.28 per 1M token |
| Confidence | High |
| Vendor country | China |
| Vendor type | Open lab |
Pricing & hardware
$0.14 / $0.28 per 1M token
This model can be self-hosted or run locally. Approximate GPU memory: 280–400 GB. Approximate system RAM: 320–640 GB. API access is available.