AI inference provider Boundless has introduced Defensive Access, a new program that gives vetted security teams access to low-refusal models tuned for security research. The first release is the DeepSeek V4.1 Flash 1017 security research version. Boundless said general-purpose models often reject legitimate testing requests because they resemble attack patterns, leaving defenders behind attackers in terms of tooling. The company said access will be granted through case-by-case review and limited to authorized security research, red-team exercises, and penetration testing. According to Vercel AI Gateway data, the model delivers throughput of 349.90 tokens per second on Boundless, a level the company says can support round-the-clock security scanning and alert investigations.
AI inference provider Boundless has launched Defensive Access, a program that offers vetted security teams access to low-refusal models tuned for security research. The initial release is the DeepSeek V4.1 Flash 1017 security research version.
Boundless said general-purpose models often reject legitimate testing requests because those prompts can resemble attack behavior. In its view, that leaves defenders behind attackers on tooling. Access under the new program is reviewed case by case and is limited to authorized security research, red-team exercises, and penetration testing.
Data from Vercel AI Gateway shows the model reaches throughput of 349.90 tokens per second on Boundless, which the company said is enough to support 24/7 security scanning and alert investigations.
This article was originally published by Bit.Fan. For more cryptocurrency news and market insights, visit www.bit.fan. Disclaimer:
The market information, project data, and third-party content displayed on this platform are for industry information sharing only and do not constitute any form of investment advice or return commitment.
Cryptocurrency trading carries high risks. Users should fully assess their risk tolerance and make independent decisions. All profits, losses, and legal responsibilities are borne by the users themselves.