DeepSeek V4 Flash is the low-cost V4 route with 1M context, thinking and non-thinking modes, tool calls and JSON output.
A compact summary of this model's best-fit workloads, routing status and production integration shape.
Model design signal
Recent public usage volume
Composite coding and reasoning signal
Gateway reports this model ready for traffic
DeepSeek policies apply for retention, training and regional handling.
enable_thinkingmax_tokenstemperaturetop_pseedresponse_formattoolsEtherGate routes requests to configured providers that can handle your prompt size and parameters, with fallbacks to maximize uptime. uptime.
Compare different providers across EtherGate
Actual cost per million tokens across providers over the past hour
| Provider | Input Price | Output Price | Cache Hit Rate |
|---|---|---|---|
| DeepSeek | ¥1.00 / MTok | ¥2.00 / MTok | 42% |
Performance metrics and benchmarks
Composite coding and reasoning signal
53.0% Win · 170.0s Avg
50.0% Win · 152.0s Avg
53.4% Win · 254.7s Avg
53.3% Win · 236.9s Avg
54.8% Win · 270.4s Avg
46.7% Win · 116.4s Avg
55.2% Win · 133.4s Avg
54.5% Win · 252.7s Avg
51.8% Win · 306.2s Avg
Top public apps this month
Available: Gateway reports this model ready for traffic
Open-source agent workspace for model routing
AI coding agent connected to EtherGate
Applied research lab for frontier models
Developer tool using OpenAI-compatible APIs
VS Code extension for model-assisted coding
Total usage per day on EtherGate
Uptime stats for DeepSeek V4 Flash across configured providers
EtherGate normalizes requests and responses across providers for you.
EtherGate supports reasoning-enabled models that can return step-by-step reasoning metadata. Use the reasoning parameter when a provider supports it, and preserve reasoning_details when continuing a conversation. Learn more about reasoning tokens.
Use the OpenAI-compatible Chat Completions endpoint. Keep your SDK, change the base URL, API key and model ID.
import OpenAI from "openai";
const client = new OpenAI({
apiKey: process.env.ETHERGATE_API_KEY,
baseURL: "https://ethergate.cn/v1",
});
const completion = await client.chat.completions.create({
model: "deepseek/deepseek-v4-flash",
messages: [{ role: "user", content: "Explain your best use case." }],
});Use this model ID directly in agent tools that accept OpenAI-compatible endpoints.
curl -fsSL 'https://ethergate.cn/console/skills/ethergate-provider/install.sh' | bash -s -- --target qclaw --api-key '<your-ethergate-key>' --locale 'en'Bearer $ETHERGATE_API_KEYContent-Type: application/jsonhttps://ethergate.cn/v1Optional attribution headers help you identify traffic by app or website when debugging usage.