DeepSeek models: prices and providers
DeepSeek has 47 models on AIModelFind (47 text). 34 have open weights, and 12 providers serve them.
All DeepSeek models
Most searched first. The price is the lowest across providers.
| Model | Lowest price | Providers | Context | License | Released |
|---|---|---|---|---|---|
| DeepSeek V4.1 Flash | $0.03/1M | 39 | 1.05M | MIT | Sep 10, 2026 |
| DeepSeek V4 Flash | $0.17/1M | 26 | 1.05M | MIT | Apr 22, 2026 |
| DeepSeek V4 Flash Vision Exp | $0.294/1M | 26 | 1.05M | MIT | Oct 17, 2026 |
| DeepSeek V4 Pro | $0.418/1M | 33 | 1.05M | MIT | Apr 22, 2026 |
| DeepSeek R1 | $1.79/1M | 9 | 164K | MIT | Jan 20, 2025 |
| DeepSeek V3 | $0.89/1M | 6 | 64K | — | Dec 25, 2024 |
| DeepSeek V4 Flash 0731 | $0.133/1M | 25 | 1.05M | MIT | Jul 31, 2026 |
| DeepSeek R1 Distill Qwen 7B | $0.01/1M | 2 | — | MIT | Jan 20, 2025 |
| DeepSeek R1 Distill Qwen 32B | $0.20/1M | 1 | — | MIT | Jan 20, 2025 |
| DeepSeek Coder V2 Instruct | $0.32/1M | 1 | — | deepseek-license | Jun 14, 2024 |
| DeepSeek Chat | $0.28/1M | 4 | 164K | — | Dec 26, 2024 |
| DeepSeek V3 0324 | $1.14/1M | 1 | 164K | MIT | Mar 24, 2025 |
| DeepSeek Coder 6.7B Instruct | Free/1M | 1 | — | deepseek | Oct 29, 2023 |
| DeepSeek Llm 67B Chat | $0.16/1M | 2 | — | deepseek | Nov 29, 2023 |
| DeepSeek-OCR | $0.02/1M | 4 | 8K | MIT | Oct 17, 2025 |
| DeepSeek OCR 2 | $0.03/1M | 2 | — | Apache 2.0 | Jan 27, 2026 |
| DeepSeek Prover V2 671B | $0.10/1M | 1 | — | — | Apr 30, 2025 |
| DeepSeek R1 0528 | $1.23/1M | 8 | 164K | MIT | May 28, 2025 |
| DeepSeek R1 0528 Qwen3 8B | $0.09/1M | 1 | — | MIT | May 29, 2025 |
| DeepSeek R1 Distill Llama 70B | $0.13/1M | 3 | 32K | MIT | Jan 20, 2025 |
| DeepSeek R1 Distill Llama 8B | $0.01/1M | 2 | — | MIT | Jan 20, 2025 |
| DeepSeek R1 Distill Qianfan Llama 8B | $0.548/1M | 1 | — | — | — |
| DeepSeek R1 Distill Qwen 1.5B | $0.01/1M | 1 | — | MIT | Jan 20, 2025 |
| DeepSeek R1 Distill Qwen 14B | $0.10/1M | 2 | — | MIT | Jan 20, 2025 |
| DeepSeek R1 Zero | $2.20/1M | 1 | — | MIT | Jan 20, 2025 |
| DeepSeek R1t Chimera | $0.02/1M | 1 | — | — | — |
| DeepSeek Reasoner | $0.28/1M | 3 | 1.05M | — | Jan 20, 2025 |
| DeepSeek TNG R1T2 Chimera | $1.20/1M | 1 | — | — | — |
| DeepSeek V2 Chat | $0.32/1M | 1 | — | deepseek | Apr 22, 2024 |
| DeepSeek V2p5 | $0.32/1M | 1 | — | deepseek | Sep 5, 2024 |
| DeepSeek V3 0324 | $0.822/1M | 7 | 128K | MIT | Mar 24, 2025 |
| DeepSeek V3.1 Terminus | $0.95/1M | 6 | 164K | MIT | Sep 22, 2025 |
| DeepSeek-V3.2-Exp | $0.308/1M | 7 | 164K | MIT | Sep 29, 2025 |
| DeepSeek-V3.1 | $0.798/1M | 15 | 164K | MIT | Aug 21, 2025 |
| Deepseek 3.2 | $0.308/1M | 25 | 164K | MIT | Dec 1, 2025 |
| DeepSeek-V3.2-Speciale | $1.68/1M | 2 | — | MIT | Nov 28, 2025 |
| DeepSeek V4.1 Flash Economy 261017 | $0.48/1M | 1 | — | — | — |
| DeepSeek V4 Flash 0423 (Thinking) | $0.51/1M | 1 | 1M | — | Apr 23, 2026 |
| DeepSeek V4 Flash 0424 | $0.20/1M | 1 | 1.05M | — | Apr 24, 2026 |
| DeepSeek V4 Flash 0731 | $1.32/1M | 1 | 1M | — | — |
| DeepSeek V4 Flash Vision Exp | $0.48/1M | 15 | 1.05M | MIT | Aug 31, 2026 |
| DeepSeek V4 Pro 0423 | $2.60/1M | 2 | 1.05M | — | Apr 23, 2026 |
| DeepSeek V4 Pro | $1.58/1M | 27 | 1M | MIT | Apr 22, 2026 |
| DeepSeek V4 Pro | $4.80/1M | 1 | 1M | — | — |
| DeepSeek V4 Pro | $3.30/1M | 1 | 1M | — | — |
| DeepSeek V4 Pro | $4.80/1M | 1 | 1M | — | — |
| DeepSeek Vl2 | $0.16/1M | 1 | — | deepseek | Dec 13, 2024 |
Where to use DeepSeek models
Providers serving the most DeepSeek models.
DeepSeek at a glance
47 models in total: 47 text. The cheapest is DeepSeek R1 Distill Qwen 7B at $0.01 /1M. DeepSeek V4.1 Flash has the largest context window, 1.05M tokens. The most searched is DeepSeek V4.1 Flash, with about 138K Google searches a month. The newest is DeepSeek V4 Flash Vision Exp, released Oct 17, 2026.
Cheapest
- DeepSeek R1 Distill Qwen 7B$0.01/1M
- DeepSeek R1 Distill Llama 8B$0.01/1M
- DeepSeek R1 Distill Qwen 1.5B$0.01/1M
- DeepSeek-OCR$0.02/1M
- DeepSeek R1t Chimera$0.02/1M
Largest context
Most searched
Newest
- DeepSeek V4 Flash Vision ExpOct 17, 2026
- DeepSeek V4.1 FlashSep 10, 2026
- DeepSeek V4 Flash Vision ExpAug 31, 2026
- DeepSeek V4 Flash 0731Jul 31, 2026
- DeepSeek V4 Flash 0424Apr 24, 2026
DeepSeek models FAQ
We track 47 DeepSeek models: Text 47.
Among DeepSeek's text models, DeepSeek R1 Distill Qwen 7B is the cheapest at $0.01 per 1M output tokens.
Yes, 34 DeepSeek models have weights you can download from DeepSeek's Hugging Face page.
12 providers serve DeepSeek models, including AIHubMix, Hugging Face, Requesty, OpenRouter, Deep Infra.
Can't find what you're looking for? Contact us