Free tools
Tools
Standalone tools backed by the llm-speed dataset. No signup, no paywall. Inspect each tool’s assumptions and available source measurements.
- Will it run — and how fast?/tools/vram-fit
Estimate model and context memory for a GPU or Mac, inspect tested model files, and compare available source benchmarks. An estimate does not guarantee a fit.
- LLM speed estimator/tools/predict
Look up measured tokens per second, separated by model size, workload and reported settings. Inspect every source run; observed ranges are not speed guarantees.