DeepSeek V3.2 pricing & benchmarks
Released Dec 1, 2025 · served by 14 providers · ranked #50 by intelligence in our index
What DeepSeek V3.2 is
DeepSeek-V3.2 is a large language model designed to harmonize high computational efficiency with strong reasoning and agentic tool-use performance. It introduces DeepSeek Sparse Attention (DSA), a fine-grained sparse attention mechanism...
- Model ID:
deepseek/deepseek-v3.2 - Modality: text->text
- Tool calling: supported · Structured output: supported
- Weights:
deepseek-ai/DeepSeek-V3.2on Hugging Face
Providers and prices for DeepSeek V3.2
The same model costs different amounts depending on who serves it. Prices are per 1M tokens, cheapest first. Uptime and throughput are OpenRouter's rolling measurements, not vendor claims.
| Provider | Input | Output | Context | Quant | Uptime 24h | Throughput |
|---|---|---|---|---|---|---|
| Baiducheapest | $0.207 | $0.311 | 131K | fp8 | 99.5% | — |
| StreamLake | $0.214 | $0.322 | 128K | fp8 | 99.8% | — |
| DeepInfra | $0.260 | $0.380 | 164K | fp4 | 98.4% | — |
| AtlasCloud | $0.260 | $0.380 | 164K | fp8 | 99.8% | — |
| Novita | $0.269 | $0.400 | 164K | fp8 | 99.9% | — |
| SiliconFlow | $0.259 | $0.420 | 164K | fp8 | 89.6% | — |
| GMICloud | $0.290 | $0.430 | 164K | fp8 | 97.6% | — |
| Venice | $0.330 | $0.480 | 160K | — | 94.1% | — |
| Phala | $1.00 | $1.00 | 164K | — | 95.7% | — |
| Alibaba | $0.370 | $1.11 | 131K | — | 97.7% | — |
| DigitalOcean | $0.425 | $1.36 | 164K | — | 99.1% | — |
| Friendli | $0.500 | $1.50 | 164K | — | 99.9% | — |
| $0.560 | $1.68 | 164K | — | 99.6% | — | |
| SambaNova | $3.00 | $4.50 | 33K | — | 98.9% | — |
DeepSeek V3.2 benchmark results
Design Arena head-to-head results, as reported through the OpenRouter model API.
| Category | Arena | Rank | Elo | Win rate |
|---|---|---|---|---|
| 3d | models | #49 | 1190 | 49.5% |
| asciiart | models | #49 | 1120 | 40.5% |
| website | models | #55 | 1200 | 50.2% |
| dataviz | models | #56 | 1187 | 48.1% |
| uicomponent | models | #56 | 1185 | 46.8% |
| codecategories | models | #57 | 1194 | 49.3% |
| gamedev | models | #57 | 1188 | 46.6% |
| svg | models | #57 | 1082 | 40.8% |
Cheaper models in the same class
Models scoring within 4 points of DeepSeek V3.2 on the Intelligence Index, but with a lower output price.
FAQ
How much does the DeepSeek V3.2 API cost?
$0.207 per 1M input tokens and $0.311 per 1M output tokens. A typical workload of 1M input + 300K output tokens costs about $0.30.
Is DeepSeek V3.2 free?
No. It is a paid model starting at $0.207 per 1M input tokens, though some providers offer trial credits.
Which provider is cheapest for DeepSeek V3.2?
Baidu at $0.207 input / $0.311 output per 1M tokens (fp8 quantization).
Can I self-host DeepSeek V3.2?
Yes — weights are published as deepseek-ai/DeepSeek-V3.2 on Hugging Face, so you can run it on your own hardware.