DeepSeek

Chinese AI company developing state-of-the-art large language models including the DeepSeek-V3 series with mixture-of-experts architecture and hybrid thinking/non-thinking capabilities

Models

Model Released Context Modalities
DeepSeek-V4-Flash-Max Apr 23, 2026 text
DeepSeek-V4-Pro-Max Apr 23, 2026 text
DeepSeek-V3.2 (Non-thinking) Dec 1, 2025 text
DeepSeek-V3.2 (Thinking) Dec 1, 2025 text
DeepSeek-V3.2 Dec 1, 2025 text
DeepSeek-V3.2-Speciale Dec 1, 2025 text
DeepSeek-V3.2-Exp Sep 29, 2025 text
DeepSeek-R1-0528 May 28, 2025 text
DeepSeek-V3 0324 Mar 25, 2025 text
DeepSeek-R1 Jan 20, 2025 text
DeepSeek R1 Distill Llama 70B Jan 20, 2025 text
DeepSeek R1 Distill Llama 8B Jan 20, 2025 text
DeepSeek R1 Distill Qwen 1.5B Jan 20, 2025 text
DeepSeek R1 Distill Qwen 14B Jan 20, 2025 text
DeepSeek R1 Distill Qwen 32B Jan 20, 2025 text
DeepSeek R1 Distill Qwen 7B Jan 20, 2025 text
DeepSeek R1 Zero Jan 20, 2025 text
DeepSeek-V3.1 Jan 10, 2025 text
DeepSeek-V3 Dec 25, 2024 text
DeepSeek VL2 Dec 13, 2024 image, text
DeepSeek VL2 Small Dec 13, 2024 text
DeepSeek VL2 Tiny Dec 13, 2024 text
DeepSeek-V2.5 May 8, 2024 text