Sunday, August 2, 2026
DeepSeek releases V4-Flash update with major agentic gains
DeepSeek published DeepSeek-V4-Flash-0731 on Hugging Face and moved its API into public beta on July 31, 2026, an MIT-licensed mixture-of-experts model with roughly 284 billion total and 13 billion active parameters. The update posted large benchmark jumps over DeepSeek's own V4-Pro despite far fewer active parameters, including Terminal-Bench 2.1 rising from 61.8 to 82.7, and the API now supports the Responses format used by coding tools like Codex.
/ Sources
/ About this story
Compiled by Venture Atlas from the sources above, using automated AI-assisted research. This is a summary of reporting published elsewhere, not original reporting - follow the source links for the full account. See our editorial standards.
Something wrong here? Email flightatlas.contact@gmail.com and we will fix it.
/ Related
- DeepSeek launches V4.1-Flash and retires V4-ProMonday, September 14, 2026
- CISA warns of industrial-scale China AI distillationWednesday, September 9, 2026
- DeepSeek taps CITIC for Shanghai STAR Market IPOWednesday, September 9, 2026
- DeepSeek plans 160,000-chip Huawei order for new data centerFriday, September 4, 2026
