DeepSeek V4 Pro
Rigorous math, proofs, and hard algorithmic problems on a budget.
What are DeepSeek V4 Pro's specs and price?
DeepSeek V4 Pro, built by DeepSeek, ships a 128K-token context window and a 64K-token max output, released 2026-05. It supports text input with a dedicated reasoning mode and costs $0.54 per million blended tokens, the 11th-cheapest of 32 models we track.
Specs
| Context window | 128K tokens |
| Max output | 64K tokens |
| Modalities | text |
| Extended thinking | Yes |
| Released | 2026-05 |
| Knowledge cutoff | 2026-02 |
| Provider | DeepSeek |
Verified 2026-08-08 — source.
Where it ranks
Strengths
- Thinking mode with visible chain-of-thought
- Frontier-level math and competition coding
- Still far cheaper than closed frontier models
More about DeepSeek V4 Pro
FAQ
What is DeepSeek V4 Pro's context window?
DeepSeek V4 Pro has a 128K-token context window and a 64K-token max output — the 31st-largest context of the 32 current models we track. Source: https://api-docs.deepseek.com/quick_start/pricing, verified 2026-08-08.
Does DeepSeek V4 Pro support vision or audio input?
No — DeepSeek V4 Pro is text-only as of 2026-08-08.
Does DeepSeek V4 Pro have a reasoning or extended-thinking mode?
Yes — DeepSeek V4 Pro exposes a dedicated reasoning mode for multi-step problems.
When was DeepSeek V4 Pro released, and what is its knowledge cutoff?
DeepSeek V4 Pro was released 2026-05 with a knowledge cutoff of 2026-02.
How much does DeepSeek V4 Pro cost, and who provides it?
DeepSeek V4 Pro is served by DeepSeek at $0.54/M blended tokens (3:1 input:output) — the 11th-cheapest of 32 current models. Full pricing breakdown: /llm-api-pricing/deepseek-v4-pro.