DeepSeek
31 models · 4 papers in the hub
Chinese lab known for strong, cost-efficient open-weight reasoning and mixture-of-experts models (DeepSeek-V3, R1).
31Models
4Papers
68.1Avg index
26Open weights
2023Active since
April 24, 2026Latest
Leading model
DeepSeek V3.2 Speciale
Models
- DeepSeek V3.2 Speciale164K$0.2989.9
- DeepSeek-V4-Flash1M$0.1089.4
- DeepSeek-V4-Pro1M$0.4488.2
- DeepSeek-V3.2131K$0.2587.1
- DeepSeek V3.1 Terminus164K$0.2783.5
- DeepSeek-R1128K$0.5575
- DeepSeek VL2129K$9.5074.9
- DeepSeek-Coder-V2$0.0074.3
- DeepSeek VL2 Small73.1
- DeepSeek V3.2 Exp164K$0.2772.2
- DeepSeek R1 Zero71.5
- DeepSeek R1 Distill Llama 70B128K$0.1070.1
- DeepSeek R1 Distill Qwen 32B128K$0.1268.4
- DeepSeek VL2 Tiny67.2
- DeepSeek R1 0528 Qwen3 8B$0.0066.2
- DeepSeek-V3 0324164K$0.2865.9
- DeepSeek R1 Distill Qwen 14B$0.0065.7
- DeepSeek-V2.58K$0.1463.4
- DeepSeek-R1-0528131K$0.5563.3
- DeepSeek-V3.1164K$0.2159.8
- DeepSeek R1 Distill Qwen 7B58.3
- DeepSeek-V3131K$0.2358.1
- DeepSeek R1 Distill Llama 8B$0.0053.3
- DeepSeek R1 Distill Qwen 1.5B$0.0032.6
- DeepSeek Coder V2 Lite Instruct$0.0030.2
- R1 Distill Qwen 32B128K$0.29
- R1 Distill Llama 70B131K$0.70
- R1164K$0.70
- DeepSeek-V2-Chat$0.00
- DeepSeek-V2128K
- DeepSeek LLM 67B Chat$0.00
Papers
- DeepSeek-V3.2: Pushing the Frontier of Open Large Language ModelsDecember 2, 2025
- DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement LearningJanuary 22, 2025
- DeepSeek-V3 Technical ReportDecember 27, 2024
- DeepSeek-V2: A Strong, Economical, and Efficient Mixture-of-Experts ModelMay 7, 2024