The Benchmark Room

This category measures time to first token, total latency, token usage, cost per successful task, schema failure rate, retries, tool-call accuracy, concurrency, and regression testing for Mistral Large 4 deployments.