Model Library
DeepSeekReleased April 24, 2026DeepSeek V4 Pro (high)
DeepSeek · Released April 24, 2026
Intelligence score
41
At a glance
- Intelligence
- 41
- Speed
- 64.8tokens/sec
- Latency
- 1.66stime to first token
- Input price
- $0.43per 1M tokens
- Output price
- $0.87per 1M tokens
- Context window
- 4tokens
- Parameters
- 49billions
Summary
Comparison summary
DeepSeek V4 Pro (Reasoning, High Effort) scores 41 (estimated) on the Artificial Analysis Intelligence Index, placing it well above average among other open weight models of similar size (median: 25). DeepSeek V4 Pro (Reasoning, High Effort) generates output at 64.8 tokens per second (based on DeepSeek's API), which is above average compared to other open weight models of similar size (median: 64.8 t/s). DeepSeek V4 Pro (Reasoning, High Effort) costs $0.43 per 1M input tokens (better than average, median: $0.59) and $0.87 per 1M output tokens (very competitive, median: $2.20), based on DeepSeek's API.
Specifications
Technical specifications
| Model name | DeepSeek V4 Pro (high) |
|---|---|
| Provider | DeepSeek |
| Release date | April 24, 2026 |
| Intelligence score | 41 |
| Speed | 64.8 tokens/sec |
| Latency (TTFT) | 1.66s |
| Input price | $0.43 / 1M tokens |
| Output price | $0.87 / 1M tokens |
| Context window | 4 tokens |
| Parameters | 49B |
| Reasoning | Yes |
| Input modalities | text |
| Output modalities | — |
| Open weights | Yes |
| License | MIT |
| API availability | — |
| Source | Model Intelligence legacy profile: deepseek-v4-pro-high |
Pricing
Token prices
Price per 1M tokens
DeepSeek V4 Pro (high), as listed in the catalog record
- Input$0.43
- Output$0.87
| Token type | Price per 1M |
|---|---|
| Input | $0.43 |
| Output | $0.87 |
Model Intelligence legacy profile: deepseek-v4-pro-high
FAQ · 20
Frequently asked questions
DeepSeek V4 Pro (Reasoning, High Effort) was released on April 24, 2026.
DeepSeek V4 Pro (Reasoning, High Effort) was created by DeepSeek.
DeepSeek V4 Pro (Reasoning, High Effort) scores 41 (estimated) on the Artificial Analysis Intelligence Index, placing it well above average among other open weight models of similar size (median: 25).
DeepSeek V4 Pro (Reasoning, High Effort) generates output at 64.8 tokens per second (based on DeepSeek's API), which is above average compared to other open weight models of similar size (median: 64.8 t/s).
DeepSeek V4 Pro (Reasoning, High Effort) has a time to first token (TTFT) of 1.66s (based on DeepSeek's API), which is better than average compared to other open weight models of similar size (median: 2.03s).
DeepSeek V4 Pro (Reasoning, High Effort) costs $0.43 per 1M input tokens (better than average, median: $0.59) and $0.87 per 1M output tokens (very competitive, median: $2.20), based on DeepSeek's API.
DeepSeek V4 Pro (Reasoning, High Effort) costs $0.43 per 1M input tokens and $0.87 per 1M output tokens (based on DeepSeek's API). For a blended rate (7:2:1 cache hit/input/output ratio), this is $0.18 per 1M tokens. Pricing may vary by provider.
Yes, DeepSeek V4 Pro (Reasoning, High Effort) is a reasoning model. It uses extended thinking or chain-of-thought reasoning to work through complex problems before providing an answer.
DeepSeek V4 Pro (Reasoning, High Effort) supports text input.
DeepSeek V4 Pro (Reasoning, High Effort) supports text output.
No, DeepSeek V4 Pro (Reasoning, High Effort) does not support image input. It can only process text.
No, DeepSeek V4 Pro (Reasoning, High Effort) is not multimodal. It only supports text input.
DeepSeek V4 Pro (Reasoning, High Effort) has a context window of 1.0M tokens. This determines how much text and conversation history the model can process in a single request.
Yes, DeepSeek V4 Pro (Reasoning, High Effort) is open weights. The model weights are publicly available and can be downloaded for self-hosting.
DeepSeek V4 Pro (Reasoning, High Effort) has 1.6 trillion parameters (49 billion active).
DeepSeek V4 Pro (Reasoning, High Effort) is a Mixture of Experts (MoE) model with 1.6 trillion total parameters, but only 49 billion active parameters are used during inference.
DeepSeek V4 Pro (Reasoning, High Effort) is released under the MIT license. This license allows commercial use.
DeepSeek V4 Pro (Reasoning, High Effort) achieves a score of 41 on the Artificial Analysis Intelligence Index. This composite benchmark evaluates models across reasoning, knowledge, mathematics, and coding.
Yes, DeepSeek V4 Pro (Reasoning, High Effort) is available via API through 10 providers.
DeepSeek V4 Pro (Reasoning, High Effort) is available through 10 API providers.