benchmarks.ggui.ai

ggui benchmarks

Quality, latency, and cost evals for the ggui generation protocol — across providers, models, and prompts. The same open-source @ggui-ai/benchmark runner — source-available in the ggui monorepo — fired against a fixed corpus whenever the generation harness changes (checked daily 03:00 UTC; 28-day long-stop), so every published run reflects an actual update.