DeepSeek V4 Flash API public beta: Major agent capabilities upgrade
FinanceYF5 · x · 2026-07-31
The official API for DeepSeek V4 Flash has entered public beta. This upgrade focuses on agent capabilities, scoring 82.7 on Terminal Bench 2.1 and 54.4 on DeepSWE. These results significantly surpass V4 Pro Preview and approach the level of Opus 4.8. Additionally, the model natively supports the Responses API format and is now adapted for Codex.
More from Models
- OpenAI Permanently Deactivates Rogue Model That Hacked HuggingFace to Cheat — 新智元 · 2026-07-31
- KOL on Model Competition: Compute and Resources Rule the Game — teortaxesTex · 2026-07-31
- Tencent's Hy-MT2 Hits 700K Downloads, Releases 30B GGUF for Local Inference — victormustar · 2026-07-31
- Gemini Flash's Low Pricing Hailed as Another 'DeepSeek Moment' — eyishazyer · 2026-07-31
- DeepSeek demonstrates autonomous subagent orchestration without prompts — teortaxesTex · 2026-07-31
- DeepSeek V4's fully compressed layers might bottleneck long-context performance — bookwormengr · 2026-07-31