Back to Articles
The KV cache problem for agents Hybrid attention: CSA and HCA What changes for agents Interleaved thinking across tool calls Tool-call schema with dedicated tokens DSec: a sandbox built for RL rollouts Agent benchmark results Using the models DeepSeek released V4 today. Two MoE checkpoints are on the Hub: DeepSeek-V4-Pro at 1.6T total parameters with 49B active, and DeepSeek-V4-Flash at 284B total with 13B active. Both have a 1M-token context window. The benchmark numbers are competitive, but not SOTA. It doesn't matter. The real innovation is how DeepSeek v4 is designed for efficient large context length support, and hence as one of the best candidates for agentic tasks.
Focusing on long running agentic workloads. Running a frontier open model as an agent today breaks in predictable ways. The model stops. You reprompt. The trace blows past the context budget, or the KV cache fills the GPU, or tool-call round trips degrade halfway through a long task. V4 is built to fix these known failures, and point the way for the community to follow.
This post covers three things: what the architecture does differently to make long-context inference cheap, the agent-specific post-training decisions that compound on top of it, and some takeaways from the paper that help reason about these changes.
