We burned 11.7bn tokens to find the best cyber AI model | GLM5.3 and DeepSeek are now frontier

https://www.aikido.dev/blog/ai-model-benchmarks-aug-21-2026

https://x.com/home/status/2091247243617841235

https://x.com/home/status/2091255520694874364

https://x.com/home/status/2091284321428525401