Your prompt. Real model responses.
Compare response text, elapsed request time, token usage and provider-reported cost. Review outputs yourself: RunLog does not assign an automated quality score.
Run a comparison
Requires an active Benchmark entitlement and a RunLog customer API key. Requests send your prompt to the configured model gateway and selected model providers. Model availability is checked by the service; unavailable models return an error. Manage your key · Check subscription availability
No benchmark has been run.
Measured results
Latency is elapsed server request time, not time to first token. Missing usage or cost is shown as unavailable. Single runs do not establish general model quality.
| Model | Latency (ms) | Input tokens | Output tokens | Cost (USD) | Status |
|---|