DeepSeek-V4-Flash

Direct from Azure models are a select portfolio curated for their market-differentiated capabilities:

• Secure and managed by Microsoft: Purchase and manage models directly through Azure with a single license, consistent support, and no third-party dependencies, backed by Azure's enterprise-grade infrastructure.
• Streamlined operations: Benefit from unified billing, governance, and seamless PTU portability across models hosted on Azure - all as part of one Microsoft Foundry platform.
• Future-ready flexibility: Access the latest models as they become available, and easily test, deploy, or switch between them within Microsoft Foundry; reducing integration effort.
• Cost control and optimization: Scale on demand with pay-as-you-go flexibility or reserve PTUs for predictable performance and savings.

Learn more about Direct from Azure models .

About this model

We present a preview version of the DeepSeek-V4 series, including two strong Mixture-of-Experts (MoE) language models — DeepSeek-V4-Pro with 1.6T parameters (49B activated) and DeepSeek-V4-Flash with 284B parameters (13B activated) — both supporting a context length of one million tokens.

We pre-train both models on more than 32T diverse and high-quality tokens, followed by a comprehensive post-training pipeline. The post-training features a two-stage paradigm: independent cultivation of domain-specific experts (through SFT and RL with GRPO), followed by unified model consolidation via on-policy distillation, integrating distinct proficiencies across diverse domains into a single model.

DeepSeek-V4-Pro-Max, the maximum reasoning effort mode of DeepSeek-V4-Pro, significantly advances the knowledge capabilities of open-source models, firmly establishing itself as the best open-source model available today. It achieves top-tier performance in coding benchmarks and significantly bridges the gap with leading closed-source models on reasoning and agentic tasks. Meanwhile, DeepSeek-V4-Flash-Max achieves comparable reasoning performance to the Pro version when given a larger thinking budget, though its smaller parameter scale naturally places it slightly behind on pure knowledge tasks and the most complex agentic workflows.

Key model capabilities

• Hybrid Attention Architecture: We design a hybrid attention mechanism combining Compressed Sparse Attention (CSA) and Heavily Compressed Attention (HCA) to dramatically improve long-context efficiency. In the 1M-token context setting, DeepSeek-V4-Pro requires only 27% of single-token inference FLOPs and 10% of KV cache compared with DeepSeek-V3.2.
• Manifold-Constrained Hyper-Connections (mHC): We incorporate mHC to strengthen conventional residual connections, enhancing stability of signal propagation across layers while preserving model expressivity.
• Muon Optimizer: We employ the Muon optimizer for faster convergence and greater training stability.