Model comparison
DeepSeek V4 Pro vs DeepSeek R1
Capability radar, value bars, and side-by-side posture.
DeepSeek V4 Pro
Not scored
value score
DeepSeek R1
90
value score
DeepSeek V4 Pro ctx
—
DeepSeek R1 ctx
128K
Capability radar
Value · context · multimodal · openness · speed posture
Value head-to-head
- DeepSeek R190
Attribute tape
| Field | DeepSeek V4 Pro | DeepSeek R1 |
|---|---|---|
| Developer | ~deepseek | ~deepseek |
| Context | — | 128K |
| Modalities | Text, Code | |
| Openness | — | Open Weights |
| Speed | — | — |
| Price | — | — |
| Value score | — | 90 |
| Summary | DeepSeek V4 Pro is a Mixture-of-Experts large language model with 1.6 trillion total parameters (49B active) and a 1 million-token context window, released as an open-weight model under MIT license. It targets advanced reasoning, long-context document processing, and agentic coding workflows. | Reasoning-focused open-weights DeepSeek model. |