DeepSeek-V3.1: API pricing, providers and specs
DeepSeek-V3.1 is a text model by DeepSeek. 15 providers serve it, from $0.798 to $4.94 /1M; the lowest price is at Poe.
How DeepSeek-V3.1 compares
Against the 737 text models on the site priced the same way.
- Mid-priced: its lowest price is $0.798 /1M against a median of $1.10; 45% of comparable models cost less.
- Its 164K-token context window is larger than 30% of them (median 262K), about 250 pages of text.
- On LMArena's Text board it ranks #26 of the 120 comparable models scored there.
- 15 providers serve it; 84% of comparable models have fewer.
- Its weights are open under the MIT license: commercial use allowed.
DeepSeek-V3.1 prices by provider
Every provider's list price, read from its own price page or API. A price marked “from” changes with context length, resolution or mode.
| Provider | Input | Output | Prompt caching | Context |
|---|---|---|---|---|
 Poe | $0.212/1M | $0.798/1M | — | 164K |
 Hugging Face | $0.25/1M | From$0.95/1M | — | — |
| Vercel AI Gateway | $0.25/1M | $0.95/1M | $0.13/1M | 164K |
 Deep Infra | $0.25/1M | From$0.95/1M | $0.13/1M | — |
| OpenRouter | $0.25/1M | From$0.95/1M | $0.13/1M | 164K |
 FastRouter | $0.25/1M | $0.95/1M | $0.13/1M | 164K |
 SiliconFlow | $0.27/1M | $1/1M | — | — |
 Jiekou.AI | $0.27/1M | $1/1M | — | — |
 Requesty | $0.30/1M | From$1/1M | $0.30/1M | 164K |
 Amazon Bedrock | $0.58/1M | From$1.68/1M | — | — |
 AIHubMix | $0.56/1M | $1.68/1M | — | 128K |
 Vertex | $0.60/1M | From$1.70/1M | $0.06/1M | — |
 Prime Inference | $0.60/1M | $1.70/1M | — | — |
| Replicate | $0.672/1M | $2.02/1M | — | — |
 Azure | $1.23/1M | From$4.94/1M | — | — |
- Poe$0.798
- Vercel AI Gateway$0.95
- FastRouter$0.95
- Hugging FaceFrom$0.95
- Deep InfraFrom$0.95
- OpenRouterFrom$0.95
- SiliconFlow$1
- Jiekou.AI$1
- RequestyFrom$1
- AIHubMix$1.68
- Amazon BedrockFrom$1.68
- Prime Inference$1.70
Specs
- Type
- Text
- Maker
- DeepSeek
- Inputs
- Text
- Outputs
- Text
- Context
- 164K≈ 250 pages of text
- Max output
- 33K
- Capabilities
- Prompt caching, JSON, Reasoning, Tool calling
- Released
- Aug 21, 2025
- License
- MITcommercial use allowed
- Open weights
- Yes, downloadable
- Parameters
- 685B
- Arena score
- 1420Text
Alternatives to DeepSeek-V3.1
Same kind of model from other makers, at a similar price.
| Model | Lowest price | Context | Arena | Providers |
|---|---|---|---|---|
| DeepSeek-V3.1this model | $0.798/1M | 164K | 1420 | 15 |
 GLM-5.3 | $1.10/1M | 1.05M | 1623 | 49 |
 Mistral Large 4 | $2.09/1M | 1.05M | 1543 | 13 |
 Qwen3.8 27B | $0.50/1M | 262K | 1591 | 33 |
| GPT-6 Luna | $0.40/1M | 1.05M | 1578 | 28 |
 Sonar | $1/1M | 127K | — | 9 |
| GLM-5.2 | $1.80/1M | 1M | 1604 | 44 |
| Grok Build 0.1 | $1.60/1M | 256K | — | 18 |
| Claude Haiku 5.5 | $0.30/1M | 1M | 1585 | 27 |
More from DeepSeek
DeepSeek V4.1 FlashDeepSeek$0.03/1MDeepSeek V4 FlashDeepSeek$0.17/1MDeepSeek V4 Flash Vision ExpDeepSeek$0.294/1MDeepSeek V4 ProDeepSeek$0.418/1MDeepSeek R1DeepSeek$1.79/1MDeepSeek V3DeepSeek$0.89/1MDeepSeek V4 Flash 0731DeepSeek$0.133/1MDeepSeek R1 Distill Qwen 7BDeepSeek$0.01/1MDeepSeek-V3.1 FAQ
The lowest price for DeepSeek-V3.1 is $0.798 /1M, at Poe.
Poe, at $0.798 /1M. 15 providers serve DeepSeek-V3.1; the table above lists every one with its price.
DeepSeek-V3.1 reads up to 164K tokens per request and writes up to 33K.
Yes. DeepSeek-V3.1's weights can be downloaded from its maker's Hugging Face page, under the MIT license.
About 685B, as the safetensors metadata of its official Hugging Face repo states.
15 API providers serve DeepSeek-V3.1. Compare their prices above, or open a provider to see everything it offers.
Can't find what you're looking for? Contact us