National Supercomputing Internet Launches DeepSeek-V4-Flash API with One-Click Access

National supercomputing network releases DeepSeek-V4-Flash API, offering one-click access to a model rivaling top closed-source systems in instruction following, hosted on a 100k-card compute pool.

Woofun AI reports that the national supercomputing internet has officially released the DeepSeek-V4-Flash-0731 model API and associated files. This iteration features extensive post-training that significantly boosts instruction-following capabilities, achieving benchmark performance comparable to leading closed-source models. Developers can deploy the service instantly via one-click access, eliminating manual environment configuration. The hosting platform now serves as a national resource pool for super-intelligent integrated computing power, boasting over 100,000 cards and hosting more than 1,700 open-source large models.