Model Efficiency

Nvidia's new SoL-Pi system cuts coding agent token usage by up to 49% with little change in performance by optimizing the control layer between the model and the environment.

Infrastructure • The Decoder

Fireworks Research released Ember-1, a specialized model built on Kimi K3 that delivers the same quality with 40% fewer tokens by cutting unnecessary reasoning. The company says quality held up in external benchmarks and live customer A/B tests.

AI Models • Hacker News