Providers
DeepSeek
Use DeepSeek through @anvia/openai.
Connection
| Field | Value |
|---|---|
| Anvia SDK | @anvia/openai |
| Compatibility | OpenAI-compatible endpoint |
| API URL | https://api.deepseek.com |
| Environment | DEEPSEEK_API_KEY |
| Provider docs | https://api-docs.deepseek.com/quick_start/pricing |
| Models | 4 |
Anvia Usage
This provider is listed as OpenAI-compatible. Start with @anvia/openai and the chat-completions adapter, then smoke test the exact workflow.
import { OpenAIClient } from "@anvia/openai";
const client = new OpenAIClient({
apiKey: process.env.DEEPSEEK_API_KEY,
baseUrl: "https://api.deepseek.com",
completionApi: "chat",
});
const model = client.completionModel("deepseek-chat");
Capabilities
| Capability | Value |
|---|---|
| Input modalities | text |
| Output modalities | text |
| Attachments | 2 / 4 models |
| Tools | 4 / 4 models |
| Structured output | 2 / 4 models |
| Reasoning | 3 / 4 models |
| Temperature | 4 / 4 models |
| Open weights | 4 / 4 models |
Models
| Model | Family | Input | Output | Capabilities | Limits | Cost | Updated |
|---|---|---|---|---|---|---|---|
deepseek-chatDeepSeek Chat |
deepseek | text | text | tools, temperature, open weights | context: 1000000 / output: 384000 | input: 0.14 / output: 0.28 / cache_read: 0.0028 | 2026-02-28 |
deepseek-reasonerDeepSeek Reasoner |
deepseek-thinking | text | text | tools, reasoning, temperature, open weights | context: 1000000 / output: 384000 | input: 0.14 / output: 0.28 / cache_read: 0.0028 | 2026-02-28 |
deepseek-v4-flashDeepSeek V4 Flash |
deepseek-flash | text | text | tools, schema, reasoning, temperature, open weights | context: 1000000 / output: 384000 | input: 0.14 / output: 0.28 / cache_read: 0.0028 | 2026-04-24 |
deepseek-v4-proDeepSeek V4 Pro |
deepseek-thinking | text | text | tools, schema, reasoning, temperature, open weights | context: 1000000 / output: 384000 | input: 0.435 / output: 0.87 / cache_read: 0.003625 | 2026-04-24 |
Read Gateway caveats before enabling this provider in production.
