DeepSeek
DeepSeek V4 Pro
DeepSeek V4 Pro from DeepSeek: Open MoE flagship with million-token context for coding and long agent runs.
- Context
- 1M
- Max output
- 384K
- Input / 1M tokens
- $0.435
- Output / 1M tokens
- $0.87
Accepts
- Text input
Produces
- Text output
AI model comparison
Compare pricing, context limits, supported inputs and deployment options to choose a model for your workflow.
DeepSeek V4 Pro from DeepSeek lists 1M context, 384K maximum output, $0.435 input per 1M tokens, $0.87 output per 1M tokens. Fugu Ultra from Sakana AI lists 1M context, $5.00 input per 1M tokens, $30.00 output per 1M tokens. Compare the inputs, outputs and deployment options below to see which fits your workflow.
DeepSeek
DeepSeek V4 Pro from DeepSeek: Open MoE flagship with million-token context for coding and long agent runs.
Accepts
Produces
Sakana AI
Fugu Ultra from Sakana AI: Quality-first multi-agent model for hard research, analysis, and competitions.
Accepts
Produces
At a glance
Compare the available figures below, then try both models on the work you need to do.
DeepSeek V4 Pro
Context capacity
Higher listed limit
DeepSeek V4 Pro:1M
Fugu Ultra:1M
Same value
Input price
Lower listed token rate
DeepSeek V4 Pro:$0.435
Fugu Ultra:$5.00
DeepSeek V4 Pro
Output price
Lower listed token rate
DeepSeek V4 Pro:$0.87
Fugu Ultra:$30.00
DeepSeek V4 Pro
Specifications
| Specification | DeepSeek V4 Pro | Fugu Ultra |
|---|---|---|
| Provider | DeepSeek | Sakana AI |
| Listed pricing route | DeepSeek | Sakana AI |
| Model family | deepseek-thinking | fugu |
| Accepted inputs | text | text, image |
| Produced outputs | text | text |
| Context window | 1M | 1M |
| Max output | 384K | — |
| Input price / 1M | $0.435Lower | $5.00 |
| Output price / 1M | $0.87Lower | $30.00 |
| Reasoning mode | Documented | Documented |
| Tool calling | Documented | Documented |
| Image input | Not documented | Documented |
| Audio input | Not documented | Not documented |
| Structured output | Documented | Documented |
| License | Open weights | Proprietary |
Workload guidance
Consider DeepSeek V4 Pro if you need a lower listed input rate or a lower listed output rate. DeepSeek V4 Pro offers downloadable weights; Fugu Ultra is listed as a proprietary model. Compare licence terms and operating costs before choosing a deployment. Test both on the same representative examples and judge the output against your acceptance criteria. Include tool failures, retries and the time to a completed task.
Explore each model’s features, API requirements and licence terms in the sources below.
DeepSeek V4 Pro
DeepSeek
Profile updated Oct 13, 2026
Source registry ↗Model card and weights ↗Publisher documentation ↗Fugu Ultra
Sakana AI
Profile updated Oct 13, 2026
Documentation and references ↗Search by model or provider to explore another comparison.