---
id: "deepseek/deepseek-v4-flash"
also: ["deepseek/deepseek-v4-flash-20260423", "deepseek-ai/DeepSeek-V4-Flash", "hf/deepseek-ai/DeepSeek-V4-Flash", "requesty/deepseek/deepseek-v4-flash", "vercel/deepseek/deepseek-v4-flash"]
card_usd_per_m: "0.22 in, 0.66 out, at deepseek"
cheapest_list_usd_per_m: "0.05 in, 0.14 out, at open-inference/fp8 on openrouter"
cheapest_effective_usd_per_m: "0.0203 in, 0.1387 out, at open-inference/fp8 on openrouter"
context: 1048576
max_output: 393216
input: ["text"]
output: ["text"]
released: "2026-04-24"
reasoning: ["xhigh", "high", "none"]
api: ["/api/v1/chat/completions", "/api/v1/completions", "/api/v1/messages", "/api/v1/responses"]
params: ["frequency_penalty", "include_reasoning", "logit_bias", "logprobs", "max_completion_tokens", "max_tokens", "min_p", "presence_penalty", "reasoning", "reasoning_effort", "repetition_penalty", "response_format", "seed", "stop", "structured_outputs", "temperature", "tool_choice", "tools", "top_a", "top_k", "top_logprobs", "top_p"]
endpoints: 49
gateways: "openrouter (27), vercel (14), requesty (5), hf (3)"
no_training: 36
trains: 3
retains_prompts: 12
coding: 56.2000
agentic: 22.2000
---
# DeepSeek: DeepSeek V4 Flash 0423

DeepSeek V4 Flash is an efficiency-optimized Mixture-of-Experts model from DeepSeek with 284B total parameters and 13B activated parameters, supporting a 1M-token context window. It is designed for fast inference and...