Another proof point for the open-weights thesis. From @RampLabs:
Fireworks AI's inference infrastructure successfully identified 7 high-severity vulnerabilities in Ramp Labs' backend. The tests utilized open-weight models like Kimi K2.6 and DeepSeek V4 Pro, demonstrating cost savings of approximately 5x compared to other methods. This event serves as further evidence supporting the effectiveness and economic advantages of open-weight models in security testing. AI
IMPACT Demonstrates cost-effective vulnerability detection using open-weight models, potentially influencing security testing strategies.