# DeepSeek V3.2 API

> DeepSeek-V3.2 is DeepSeek's open-weight, reasoning-first model released in December 2025 as the official successor to V3.2-Exp. It introduces DeepSeek Sparse Attention to reduce long-context training and inference compute and is DeepSeek's first model to integrate thinking directly into tool use, trained on agent data synthesized across more than 1,800 environments. It supports tool calling in both thinking and non-thinking modes, suiting agents and general reasoning.

Source: https://www.sandbase.ai/models/deepseek/deepseek-v3.2
API reference: https://www.sandbase.ai/models/deepseek/deepseek-v3.2/docs

## Facts

- Provider: DeepSeek
- Type: Language model
- Model ID: `deepseek/deepseek-v3.2`
- Price: $0.25 per 1M input tokens, $0.38 per 1M output tokens
- Context window: 131.1K tokens, up to 65.5K output tokens
- Capabilities: chat, reasoning, structured output, function calling
- Billing: pay per use from one SandBase balance; no provider account or subscription needed

## Call it

- Chat Completions API: `POST https://api.sandbase.ai/v1/chat/completions`

```bash
curl -X POST https://api.sandbase.ai/v1/chat/completions \
  -H "Authorization: Bearer $SANDBASE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
  "model": "deepseek/deepseek-v3.2",
  "messages": [
    {
      "role": "user",
      "content": "Hello"
    }
  ],
  "max_tokens": 2048
}'
```

## FAQ

### How do I call DeepSeek V3.2 through SandBase?

Create a SandBase API key, then send requests with the model ID "deepseek/deepseek-v3.2" to Chat Completions API (/v1/chat/completions) at https://api.sandbase.ai. The request examples on this page show the exact payload.

### How much does DeepSeek V3.2 cost?

$0.25 per 1M input tokens and $0.38 per 1M output tokens, billed per request from your SandBase balance. Discounts, when available, are shown in the pricing on this page.

### What is the context window of DeepSeek V3.2?

131.1K tokens of context, with up to 65.5K output tokens per response.

### How does DeepSeek V3.2 score on benchmarks?

DeepSeek V3.2 scores 21.5 on the Artificial Analysis Intelligence Index (v4.3), ranking #60 of 83 models on SandBase; Coding Index 44.2. Source: Artificial Analysis, measured as DeepSeek V3.2 (Reasoning).

### Which features does DeepSeek V3.2 support?

DeepSeek V3.2 supports reasoning, function calling, structured output. See Specifications above for the full list.

### Do I need a separate DeepSeek account?

No. One SandBase API key and balance gives you access to DeepSeek V3.2 and the other models in the catalog; you do not need to sign up with DeepSeek separately.