DeepSeek-V3.1 is a 671B-parameter hybrid reasoning model (37B active) with thinking and non-thinking modes selected through prompt templates. It extends DeepSeek-V3 with two-phase long-context training (up to 128K tokens) and uses FP8 microscaling for efficient inference. The model does well at tool use, code generation, and reasoning, with performance comparable to DeepSeek-R1 and faster responses. It supports structured tool calling, code agents, and search agents, which suits research and agentic workflows. It succeeds DeepSeek V3-0324.
$0.125market price $0.25
$0.065market price $0.13
$0.475market price $0.95
Throughput is how fast the model writes (tokens per second, higher is better). Time to first token is how long before the answer starts (lower is better).
Share of requests served without a server error. Client errors (4xx) do not count.
Works with the OpenAI SDK pointed at our base URL. Every parameter is in the API reference.
Tokens processed per day on NagaAI.
DeepSeek-V3.1 is a 671B-parameter hybrid reasoning model (37B active) with thinking and non-thinking modes selected through prompt templates. It extends DeepSeek-V3 with two-phase long-context training (up to 128K tokens) and uses FP8 microscaling for efficient inference.
On NagaAI, DeepSeek Chat v3.1 costs $0.125 per 1M input tokens and $0.475 per 1M output tokens. You pay only for what you use, with no subscription.
DeepSeek Chat v3.1 reads up to 32.8k tokens in one request, counting the prompt and the answer together.
Yes. DeepSeek Chat v3.1 can call your functions as tools and can answer in a JSON schema you give it.
It takes text as input and returns text.
Create an API key, copy the model ID from the top of this page and send requests with the OpenAI or Anthropic SDK pointed at NagaAI. The API section on this page has a ready example.
DeepSeek V4.1 Flash 1017DeepseekDeepSeek V4.1 Flash 0910DeepseekDeepSeek V4 Flash Vision ExpDeepseekDeepSeek V4 Pro 0813DeepseekDeepSeek V4 Flash 0731DeepseekDeepSeek V4 Pro 0423Deepseek