- DeepSeek released this API model on October 17, 2026. The source describes it as a cost-efficient tier aimed at coding tools and agentic tasks that carry out multiple steps on their own. It accepts text and image input and answers with text.
- Pricing is 0.10 dollars per million input tokens and 0.50 dollars per million output tokens, with cached reads at 0.002 dollars per million tokens. The context window is 1,048,576 tokens. Prices differ by provider for the same model, and DeepSeek's direct price is shown as 0.15 dollars for input and 0.60 dollars for output.
- Artificial Analysis measured results with the Max setting, reporting an overall Intelligence Index of 39.5 and a terminal task benchmark score of 26.8%. The speed and completion time advantage over the existing V4 Pro claimed in the source is not confirmed by measurements on this page.
- Three providers offer this model: inference.net, DeepInfra, and DeepSeek. The overall OpenRouter availability shown on the page is 99.91% over the last 3 days.
O: Useful for small development teams where API cost is a burden, when a developer runs the same coding agent tasks as with existing models in separate runs and compares token cost and completion time.
X: The performance advantage claim in the original article comes from DeepSeek's announcement, and the benchmark is based on the Max setting, so the speed advantage over the existing V4 Pro should not be treated as confirmed until you measure it yourself.