DeepSeek Launches V4 Flash API in Public Beta
Get the Tech newsletter
Daily tech — startups, AI labs, chips, the launches that shape the next decade. Free.
- DeepSeek rolled out the official V4 Flash API in public beta on July 31, 2026, touting enhanced agent capabilities and benchmark scores "far surpassing" V4 Pro Preview.
- Nikkei Asia reports the release arrives as the AI price war heats up, while Superintelligence notes DeepSeek's move came "overnight" in direct response to OpenAI's price cut.
- Arena reports DeepSeek-V4-Flash-High scored 1586 on the Frontend Code Arena and is priced at $0.14/$0.28 per million tokens, described as the best performance-per-dollar of any model in its class.
- DeepSeek clarified that V4-Flash-0731 keeps the exact same model architecture and size as the preview version, and that the official V4-Pro release is coming "ASAP."
- BleepingComputer reports a hacker used DeepSeek AI to autonomously attack vulnerable servers, underscoring the offensive-security stakes of the agent capabilities now exposed via the API.
- Max Weinbach demonstrated running the model at 40 tokens/second on an M3 Ultra Mac Studio with 512GB of memory, signaling strong local-inference viability.
Why it matters: DeepSeek's API posts the best price-to-performance ratio Arena has measured in its class ($0.14/$0.28 per MToken at a 1586 score) while landing within hours of OpenAI's price cut, compressing margins across the frontier-model market. The simultaneous hacker report via BleepingComputer shows those same agent capabilities are already being weaponized in the wild.
