---
id: "deepseek/deepseek-chat-v3.1"
also: ["deepseek/deepseek-chat-v3.1", "deepseek-ai/DeepSeek-V3.1", "hf/deepseek-ai/DeepSeek-V3.1"]
cheapest_list_usd_per_m: "0.25 in, 0.95 out, at deepinfra on hf"
cheapest_effective_usd_per_m: "0.1969 in, 0.9424 out, at deepinfra/fp4 on openrouter"
context: 163840
max_output: 147456
input: ["text"]
output: ["text"]
released: "2025-08-21"
api: ["/api/v1/chat/completions", "/api/v1/completions", "/api/v1/messages", "/api/v1/responses"]
params: ["frequency_penalty", "include_reasoning", "logit_bias", "logprobs", "max_tokens", "min_p", "presence_penalty", "reasoning", "repetition_penalty", "response_format", "seed", "stop", "structured_outputs", "temperature", "tool_choice", "tools", "top_k", "top_logprobs", "top_p"]
endpoints: 11
gateways: "openrouter (8), hf (3)"
no_training: 8
trains: 0
retains_prompts: 1
---
# DeepSeek: DeepSeek V3.1

DeepSeek-V3.1 is a large hybrid reasoning model (671B parameters, 37B active) that supports both thinking and non-thinking modes via prompt templates. It extends the DeepSeek-V3 base with a two-phase long-context...