DeepSeek V4 Flash
deepseek-v4-flash
These are Mengbi Router customer prices in USD per 1M tokens. Provider: DeepSeek; Context: 1M; Input price: $0.44; Output price: $1.32.
- Low latency
- Cost effective
AI models and pricing
Model specs
- Provider
- DeepSeek
- Context
- 1M
- Max output
- 16K
- Released
- Apr 2026
- Input modalities
- Text
- Output modalities
- Text
- Tool calling
- Supported
- Structured output
- Supported
- Currency
- USD
- Model ID
- deepseek-v4-flash
Model pricing and access details
USD / 1M tokens| Model | Official | Mengbi | Discount |
|---|---|---|---|
| Input price | $0.44 | $0.44 | |
| Output price | $1.32 | $1.32 | |
| Cache read price | $0.014 | $0.014 |
Estimated cost for common usage
Estimates based on current Mengbi Router prices, so you can plan real request costs.
| Token volume | Input | Output | Input + output total |
|---|---|---|---|
| 10K tokens | $0.0044 | $0.0132 | $0.0176 |
| 100K tokens | $0.044 | $0.132 | $0.176 |
| 1M tokens | $0.44 | $1.32 | $1.76 |
| 10M tokens | $4.4 | $13.2 | $17.6 |
Estimates use the current prices above and exclude cache-hit discounts. Your invoice is metered by the actual token usage of each request.
Public benchmark scores
How DeepSeek V4 Flash scores on common public benchmarks, so you can compare capability at a glance.
| Benchmark | Score |
|---|---|
| GPQA Diamond | 80.1 |
| SWE-bench Verified | 60.3 |
| AIME 2026 | 78.8 |
Curated from vendor-published evaluations (snapshot 2026-09). For reference only; not independently tested by Mengbi.
API protocol compatibility
Keep your SDK. Change one line.
Use the protocol your stack already speaks.
const client = new OpenAI({
apiKey: process.env.MENGBI_ROUTER_KEY,
baseURL: process.env.MENGBI_ROUTER_BASE_URL,
})
const response = await client.chat.completions.create({
model: 'deepseek-v4-flash',
messages,
})Frequently asked questions
Common questions about the DeepSeek V4 Flash API: pricing, capabilities, and integration.
How much does the DeepSeek V4 Flash API cost?
Through Mengbi Router, DeepSeek V4 Flash costs $0.44 per 1M input tokens and $1.32 per 1M output tokens. Cached input is billed at $0.014 per 1M tokens.
What is the context window of DeepSeek V4 Flash?
DeepSeek V4 Flash supports a 1M-token context window. A single request can generate up to about 16K output tokens.
Does DeepSeek V4 Flash support tool calling and structured output?
Yes. DeepSeek V4 Flash supports the tools parameter in chat.completions for function calling. It also supports response_format with a JSON Schema for structured output.
What input and output formats does DeepSeek V4 Flash support?
It accepts Text and returns Text.
How do I call DeepSeek V4 Flash with the OpenAI SDK?
Keep your OpenAI SDK, point the base URL at Mengbi Router, use your Mengbi API key, and set model to deepseek-v4-flash in the request.
Which company provides DeepSeek V4 Flash?
DeepSeek V4 Flash is provided by DeepSeek and available through Mengbi Router with one API key and pay-as-you-go billing. The model was first released in Apr 2026.