DeepSeek Releases V4-Flash: Major Agent Upgrades, Now Open-Source
victormustar · x · 2026-07-31
DeepSeek officially announced that the DeepSeek-V4-Flash API is now in public beta. The company stated that this version features massive upgrades in Agent capabilities, with benchmark scores far surpassing the V4-Pro-Preview.
Third-party tests show the model scored 40 on CyberGym, 50 on DeepSWE, and 20 on Terminal Bench 2.1. The model is open-weight under the MIT license, natively supports the Responses API format, and is fully adapted for Codex.
Related event: DeepSeek-V4-Flash GA Public Beta: Big Agent Capability Leap(31 posts)→
More from Models
- antirez Begins Converting DeepSeek v4 Flash to GGUF for Local Inference — antirez · 2026-07-31
- GPT 5.6 Luna Beats Google's Best in Intelligence and Undercuts Its Cheapest — Rare_Bunch4348 · 2026-07-31
- DeepSeek Costs 1/3 of GPT Luna for Coding: A Practical Token & Expense Breakdown — auto_off · 2026-07-31
- Chinese LLMs on the Rise: Matching US Frontier Models at a Fraction of the Cost — repbre · 2026-07-31
- DeepSeek-V4-Flash Repo Surfaces on Hugging Face with Million-Token Context — NielsRogge · 2026-07-31
- DeepSeek-V4-Flash Agent Eval: Completes 3D Task for $0.07 — cedric_chee · 2026-07-31