GeekRadar / The comparison desk
COMPARE.
AI models
deepseek

DeepSeek-V3 (Mar 2025)

AI model Released

Open weights
Yes*

Capability & preference

General capability · ECI

A cross-benchmark capability estimate. Index points, not a percentage; intervals show uncertainty.

DeepSeek-V3 (Mar 2025) 135.2* DeepSeek-V3 (Mar 2025) · 95% interval 133.0–137.3

Run it locally

Inference · memory estimate

DeepSeek-V3 (Mar 2025)

Not enough verified architecture information for a reliable VRAM estimate.

Capabilities

Open weights
Yes*

System configuration

Detailed specifications

Specifications · 4 fields
Access model
Open weights (restricted use)*
Open weights
Yes*
Organization
DeepSeek*
Organization country
China*

* Provisional data