Workload
endurance-30m-v1
released
Question it answers
Does this system degrade under sustained load — and when?
What is measured
Cycles the interactive-assistant-v2 human corpus closed-loop at concurrency 4 for 180 minutes; per-request start offsets give 5-minute windows with checkpoints at 30/60/180 min; windowed inter-token latency, TTFT and request success; format, language and repetition gates. The “30m” in the id is historical — the first checkpoint.
Claims this workload cannot support
Lifetime reliability claims; thermal claims without documented ambient conditions.
Frozen identity
Every released workload revision freezes its item hashes, load model and quality gates. Results reference the exact revision; changing any identity field creates a new revision. The facts below are read from the catalog file, not typed into this page.
- Revision
- 2026-08-03
- Load model
- closed_loop
- Quality gates
- format_gate, language_gate, repetition_gate
- Metrics contract
- itl_ms_by_window, ttft_ms_by_window, requests_ok_by_window
- Corpus hash
- 660eaad10d3f…