DeepSeek · Model
DeepSeek V3.2 pricing, context window and specs
DeepSeek V3.2 costs $0.259 per million input tokens and $0.80 per million output tokens, and reads up to 163,840 tokens per request. By blended price it is number 2 of DeepSeek’s 4 current models here, counting from the cheapest.
- Input / 1M
- $0.259
- Output / 1M
- $0.80
- Cached / 1M
- $0.14
- Context
- 163,840
- Max output
- Not published
- Released
- 2025-12-01
Estimate your cost with DeepSeek V3.2Compare all DeepSeek modelsCount tokens for DeepSeek V3.2
Pricing
DeepSeek V3.2 API pricing
| Price per 1M tokens | Input | Output |
|---|---|---|
| Standard | $0.259 | $0.80 |
| Cached input (read) | $0.14 | – |
Output costs 3.1× the input price, so long replies drive the bill more than long prompts. If your requests share a long fixed prefix, such as a system prompt or tool definitions, caching cuts that part to 54% of the normal input price.
History
DeepSeek V3.2 price history
Price increase, +32% blended
- Output
- from $0.42 to $0.80
- +90%
- Cached input
- from $0.135 to $0.14
- +3.7%
DeepSeek V3.2 has open weights, so this is the price of OpenRouter’s top-ranked host, which can move when hosts change rather than because DeepSeek changed its price. Every change across all models is in the AI pricing changes log.
Examples
What DeepSeek V3.2 costs in practice
| Workload | Tokens in / out | Requests/day | Per request | Per month |
|---|---|---|---|---|
| Support chatbot50% cached | 1,500 / 400 | 1,000 | $0.000619 | $18.84 |
| RAG app20% cached | 6,000 / 500 | 500 | $0.00181 | $27.55 |
| Coding agent80% cached | 40,000 / 2,000 | 200 | $0.00815 | $49.59 |
| Document summariser | 8,000 / 600 | 300 | $0.00255 | $23.29 |
Change any number in the cost calculator.
Context
Context window
DeepSeek V3.2 accepts up to 163,840 tokens per request, roughly 122,880 English words or 246 pages. That budget is shared between your prompt and the reply, and no separate cap on the reply is published.
Filling the whole window with one prompt costs about $0.0424 at list price, so for repeated long-document work, retrieval (sending only the relevant parts) or prompt caching is usually much cheaper.
Features
Capabilities
- Image input
- No
- Tool / function calling
- Yes
- Reasoning mode
- Yes
- Prompt caching
- Yes
- Structured output
- Yes
- Open weights
- Yes
- Accepts
- text
- Knowledge cutoff
- Not published
As declared by the provider’s API. It shows a feature exists, not how well it works.
API
DeepSeek V3.2 model ID
- OpenRouter
deepseek/deepseek-v3.2- Open weights
- deepseek-ai/DeepSeek-V3.2
Use the OpenRouter ID when calling it through OpenRouter; DeepSeek’s own API may name it differently. See the model ID reference for the providers we document.
DeepSeek
Where DeepSeek V3.2 sits in DeepSeek’s line-up
Ranked by blended price (three parts input to one part output, $0.3943 per 1M for DeepSeek V3.2), it is number 2 of DeepSeek’s 4 current models here, counting from the cheapest. One step down, DeepSeek V4 Pro 0423 costs $0.261 blended (34% less). Unlike DeepSeek V3.2, it reads up to 1,024,000 tokens (163,840 here). One step up, DeepSeek V4.1 Flash costs $0.525 (1.3× as much). Unlike DeepSeek V3.2, it accepts images and reads up to 1,048,576 tokens (163,840 here). See every DeepSeek model and price
| Model | Input / 1M | Output / 1M | Context | Released |
|---|---|---|---|---|
| DeepSeek V4 Pro 0423 | $0.2088 | $0.4176 | 1,024,000 | 2026-04-24 |
| DeepSeek V3.2 (this page) | $0.259 | $0.80 | 163,840 | 2025-12-01 |
| DeepSeek V4.1 Flash | $0.30 | $1.20 | 1,048,576 | 2026-09-10 |
| DeepSeek V4 Pro 0813 | $1.32 | $3.96 | 1,048,576 | 2026-08-12 |
DeepSeek’s current models with a page here, cheapest first. Older and superseded models are listed on the provider page and in the comparison table.
History
Earlier versions of DeepSeek V3.2
The previous model in this line is DeepSeek V3.2 Exp, first listed on 2025-09-29. Compared with it, DeepSeek V3.2’s input price is 4% lower and its output price is 95% higher.
| Model | Input / 1M | Output / 1M | Context | Released |
|---|---|---|---|---|
| DeepSeek V3.2 Exp | $0.27 | $0.41 | 163,840 | 2025-09-29 |
| DeepSeek V3.1 | $0.25 | $0.95 | 163,840 | 2025-08-21 |
| DeepSeek V3 0324 | $0.29 | $1.14 | 128,000 | 2025-03-24 |
| DeepSeek V3 | $0.32 | $0.89 | 163,840 | 2024-12-26 |
Earlier releases with the same name and size in our data, newest first. Dates are when our source first listed each model.
Save
Cheaper alternatives
- Nemotron 3.5 LightningNVIDIA74% cheaper
- Gemma 4 26B A4BGoogle73% cheaper
- Qwen3.5-9BQwen71% cheaper
- Seed 1.6 FlashByteDance Seed67% cheaper
- GPT-6 LunaOpenAI49% cheaper
Current models from major providers, one per provider, that keep tool calling, a reasoning mode, a usable context window and output limit. Compared by blended price.
Compare
Similarly priced models
- Qwen3 VL 8B InstructQwen$0.25 / $0.75
- GLM 4.6VZ.ai$0.30 / $0.90
- Muse Glimmer 30BMeta$0.30 / $1.20
- MiniMax M3MiniMax$0.30 / $1.20
- Mistral Small 4Mistral$0.15 / $0.60
Current models from other major providers, closest blended price.
FAQ
DeepSeek V3.2 questions
How much does DeepSeek V3.2 cost?
DeepSeek V3.2 costs $0.259 per million input tokens and $0.80 per million output tokens, with cached input at $0.14. A typical chatbot reply (1,500 tokens in, 400 out) costs about $0.000709. Use the LLM cost calculator for your own workload.
What is DeepSeek V3.2’s context window?
163,840 tokens, about 122,880 English words or 246 printed pages, shared between your prompt and the reply. No separate limit on the reply is published, so the window is the limit.
Does DeepSeek V3.2 support prompt caching and batch requests?
Yes, cached input is billed at $0.14 per million tokens (54% of the normal input price). No batch price is published for it.
What are cheaper alternatives to DeepSeek V3.2?
With the same essentials (tool calling), the cheapest options are Nemotron 3.5 Lightning (74% cheaper), Gemma 4 26B A4B (73% cheaper), Qwen3.5-9B (71% cheaper). Cheaper doesn’t mean equivalent, so test them on your own prompts.
What is the API model ID for DeepSeek V3.2?
On OpenRouter it is deepseek/deepseek-v3.2. DeepSeek’s own API may use a different name, so check its model list before you deploy. Pass the ID as the model value in each request.
What did DeepSeek V3.2 replace?
The previous DeepSeek model in the same line is DeepSeek V3.2 Exp, first listed on 2025-09-29. Compared with it, DeepSeek V3.2’s input price is 4% lower and its output price is 95% higher. Prices are only part of the picture, so test the newer model on your own prompts before switching.
Is DeepSeek V3.2 open source?
Yes, DeepSeek V3.2 has open weights, so you can run it yourself or choose between several hosting providers. The prices here are a typical hosted price.