{ "run_id": "2026-10-09-core-greedy-nothink-adb6a2", "suite": "decision-conformance", "tier": "core", "suite_hash": "554e796a858218e48bcb83da003554769782a4a32d3e061da87510d7f9cea490", "subject": "jev-27b-vl/hf-ba3f0d5/gguf-q3_k_m-iammrduncan/llamacpp-cuda-b11370-jev", "mode": "native", "profile": "greedy-nothink", "status": "complete", "label": null, "items": { "planned": 17, "attempted": 17, "succeeded": 17, "failed": 0, "failed_by_status": {} }, "metrics": { "checks": { "invariants": { "passed": 33, "total": 33, "failures": [] }, "batched_vs_separate": { "passed": 6, "total": 6, "failures": [], "max_abs_diff": 0 }, "repeat_stability": { "passed": 6, "total": 6, "failures": [], "max_std": 0 }, "option_order": { "passed": 2, "total": 2, "failures": [], "max_confidence_delta": 0.038137 }, "monotonic": { "passed": 2, "total": 2, "failures": [], "expected_scores": [ 0.644412, 1.937897, 2.999482 ] } }, "passed": 49, "total": 49 }, "latency_ms": { "p50": 6506.050917, "p95": 10600.669875, "p99": 10600.669875, "note": "successful requests, nearest rank, measured from America/Chicago" }, "usage": { "input_tokens": 4776, "output_tokens": 0, "cost_usd": 0 } }