Code & Chain · Signal Desk

DeepSeek Open-Sources V4-Flash-0731 Model with Million-Token Context

Original sourceDataNorth AI

Summary

DeepSeek released V4-Flash-0731 on July 31, a 284B parameter Mixture-of-Experts (MoE) model activating 13B parameters per token, supporting a 1,000,000 token context window. Weights are released under MIT license on Hugging Face, API offers public beta at $0.14 per 1M input tokens and $0.28 per 1M…

Key points

  • Developers get an efficient, open model for coding and agents, enabling low-cost AI application development.
  • DeepSeek's open-source model and low-cost API provide an alternative to large models like GPT-5.6, especially for agentic tasks.
  • Developers can leverage the million-token context and low-cost model for large-scale code processing and agent development, reducing AI application costs.

Editorial note

This page is Code & Chain's editorial summary of public sources. It may be prepared with AI assistance and published through an automated workflow. Refer to the original sources; this content is not investment, legal, or tax advice.