{ "caveats": [ "Finite lab cohort; configured 262144 context differs from measured inputs. Aggregate uses first-output to final-completion window. Whole-host memory sampled during speed only; missing counters remain unassessed. Broader coding quality and delivery acceptance unassessed." ], "engine": { "name": "NVIDIA NGC vLLM", "recipe": "plain", "version": "0.29.0+5013de39.dev" }, "format": "olir/0.1", "history": { "corrections": [], "supersedes": [] }, "id": "d65163d3e20a4aac92032dc2", "measured_at": "2026-10-07T04:19:07.006640+00:00", "metrics": [ { "basis": null, "n": null, "name": "wall_seconds", "scope": "machine_total", "source_field": "wall_seconds", "statistic": "single", "unit": "s", "value": 202.04653906822205 }, { "basis": null, "n": 20, "name": "ttft_median_s", "scope": "per_person", "source_field": "ttft_median_s", "statistic": "median", "unit": "s", "value": 114.3505926100188 }, { "basis": null, "n": null, "name": "needle_total", "scope": "machine_total", "source_field": "needle_total", "statistic": "single", "unit": "count", "value": 60 }, { "basis": null, "n": 20, "name": "ttft_max_s", "scope": "per_person", "source_field": "ttft_max_s", "statistic": "max", "unit": "s", "value": 162.46046369097894 }, { "basis": null, "n": null, "name": "requests_ok", "scope": "machine_total", "source_field": "requests_ok", "statistic": "single", "unit": "count", "value": 20 }, { "basis": null, "n": null, "name": "needle_correct", "scope": "machine_total", "source_field": "needle_correct", "statistic": "single", "unit": "count", "value": 60 }, { "basis": null, "n": null, "name": "preemptions", "scope": "machine_total", "source_field": "preemptions", "statistic": "single", "unit": "count", "value": 0 }, { "basis": null, "n": null, "name": "decode_overlap", "scope": "machine_total", "source_field": "decode_overlap", "statistic": "single", "unit": "count", "value": 20 }, { "basis": null, "n": null, "name": "aggregate_concurrency", "scope": "machine_total", "source_field": "aggregate_concurrency", "statistic": "single", "unit": "count", "value": 20 }, { "basis": null, "n": null, "name": "requests_total", "scope": "machine_total", "source_field": "requests_total", "statistic": "single", "unit": "count", "value": 20 }, { "basis": null, "n": null, "name": "correctness_probes_passed", "scope": "machine_total", "source_field": "correctness_probes_passed", "statistic": "single", "unit": "count", "value": 20 }, { "basis": null, "n": null, "name": "aggregate_decode_tps", "scope": "machine_total", "source_field": "aggregate_tps", "statistic": "single", "unit": "tok/s", "value": 71.51700109014168 }, { "basis": null, "n": null, "name": "aggregate_e2e_tps", "scope": "machine_total", "source_field": "aggregate_end_to_end_tps", "statistic": "single", "unit": "tok/s", "value": 50.68139274854103 }, { "basis": null, "n": null, "name": "correctness_probes_total", "scope": "machine_total", "source_field": "correctness_probes_total", "statistic": "single", "unit": "count", "value": 20 }, { "basis": null, "n": 20, "name": "decode_tps", "scope": "per_person", "source_field": "per_request_decode_tps", "statistic": "mean", "unit": "tok/s", "value": 6.893528752744044 }, { "basis": null, "n": null, "name": "forced_output_tokens", "scope": "machine_total", "source_field": "forced_output_tokens", "statistic": "single", "unit": "tokens", "value": 512 }, { "basis": null, "n": null, "name": "peak_host_unavailable_gib", "scope": "machine_total", "source_field": "peak_host_unavailable_gib", "statistic": "max", "unit": "GiB", "value": 93.79351043701172 }, { "basis": null, "n": null, "name": "ttft_s", "scope": "per_person", "source_field": "team_load_requests[0].ttft_s", "statistic": "single", "unit": "s", "value": 93.53175729000941 }, { "basis": null, "n": null, "name": "total_s", "scope": "per_person", "source_field": "team_load_requests[0].total_s", "statistic": "single", "unit": "s", "value": 199.94406931602862 }, { "basis": null, "n": null, "name": "decode_tps", "scope": "per_person", "source_field": "team_load_requests[0].decode_tps", "statistic": "single", "unit": "tok/s", "value": 4.802075908989307 }, { "basis": null, "n": null, "name": "user_number", "scope": "per_person", "source_field": "team_load_requests[0].user_number", "statistic": "single", "unit": "count", "value": 0 }, { "basis": null, "n": null, "name": "prompt_tokens", "scope": "per_person", "source_field": "team_load_requests[0].prompt_tokens", "statistic": "single", "unit": "tokens", "value": 40100 }, { "basis": null, "n": null, "name": "output_tokens", "scope": "per_person", "source_field": "team_load_requests[0].output_tokens", "statistic": "single", "unit": "tokens", "value": 512 }, { "basis": null, "n": null, "name": "needle_correct", "scope": "per_person", "source_field": "team_load_requests[0].needle_correct", "statistic": "single", "unit": "count", "value": 3 }, { "basis": null, "n": null, "name": "needle_total", "scope": "per_person", "source_field": "team_load_requests[0].needle_total", "statistic": "single", "unit": "count", "value": 3 }, { "basis": null, "n": null, "name": "ttft_s", "scope": "per_person", "source_field": "team_load_requests[1].ttft_s", "statistic": "single", "unit": "s", "value": 162.12879321700893 }, { "basis": null, "n": null, "name": "total_s", "scope": "per_person", "source_field": "team_load_requests[1].total_s", "statistic": "single", "unit": "s", "value": 202.01941937801894 }, { "basis": null, "n": null, "name": "decode_tps", "scope": "per_person", "source_field": "team_load_requests[1].decode_tps", "statistic": "single", "unit": "tok/s", "value": 12.810027045889365 }, { "basis": null, "n": null, "name": "user_number", "scope": "per_person", "source_field": "team_load_requests[1].user_number", "statistic": "single", "unit": "count", "value": 1 }, { "basis": null, "n": null, "name": "prompt_tokens", "scope": "per_person", "source_field": "team_load_requests[1].prompt_tokens", "statistic": "single", "unit": "tokens", "value": 40103 }, { "basis": null, "n": null, "name": "output_tokens", "scope": "per_person", "source_field": "team_load_requests[1].output_tokens", "statistic": "single", "unit": "tokens", "value": 512 }, { "basis": null, "n": null, "name": "needle_correct", "scope": "per_person", "source_field": "team_load_requests[1].needle_correct", "statistic": "single", "unit": "count", "value": 3 }, { "basis": null, "n": null, "name": "needle_total", "scope": "per_person", "source_field": "team_load_requests[1].needle_total", "statistic": "single", "unit": "count", "value": 3 }, { "basis": null, "n": null, "name": "ttft_s", "scope": "per_person", "source_field": "team_load_requests[2].ttft_s", "statistic": "single", "unit": "s", "value": 154.09069847100182 }, { "basis": null, "n": null, "name": "total_s", "scope": "per_person", "source_field": "team_load_requests[2].total_s", "statistic": "single", "unit": "s", "value": 201.7493742229999 }, { "basis": null, "n": null, "name": "decode_tps", "scope": "per_person", "source_field": "team_load_requests[2].decode_tps", "statistic": "single", "unit": "tok/s", "value": 10.72207718609505 }, { "basis": null, "n": null, "name": "user_number", "scope": "per_person", "source_field": "team_load_requests[2].user_number", "statistic": "single", "unit": "count", "value": 2 }, { "basis": null, "n": null, "name": "prompt_tokens", "scope": "per_person", "source_field": "team_load_requests[2].prompt_tokens", "statistic": "single", "unit": "tokens", "value": 40102 }, { "basis": null, "n": null, "name": "output_tokens", "scope": "per_person", "source_field": "team_load_requests[2].output_tokens", "statistic": "single", "unit": "tokens", "value": 512 }, { "basis": null, "n": null, "name": "needle_correct", "scope": "per_person", "source_field": "team_load_requests[2].needle_correct", "statistic": "single", "unit": "count", "value": 3 }, { "basis": null, "n": null, "name": "needle_total", "scope": "per_person", "source_field": "team_load_requests[2].needle_total", "statistic": "single", "unit": "count", "value": 3 }, { "basis": null, "n": null, "name": "ttft_s", "scope": "per_person", "source_field": "team_load_requests[3].ttft_s", "statistic": "single", "unit": "s", "value": 125.18151886499254 }, { "basis": null, "n": null, "name": "total_s", "scope": "per_person", "source_field": "team_load_requests[3].total_s", "statistic": "single", "unit": "s", "value": 200.92618343001232 }, { "basis": null, "n": null, "name": "decode_tps", "scope": "per_person", "source_field": "team_load_requests[3].decode_tps", "statistic": "single", "unit": "tok/s", "value": 6.746349765155457 }, { "basis": null, "n": null, "name": "user_number", "scope": "per_person", "source_field": "team_load_requests[3].user_number", "statistic": "single", "unit": "count", "value": 3 }, { "basis": null, "n": null, "name": "prompt_tokens", "scope": "per_person", "source_field": "team_load_requests[3].prompt_tokens", "statistic": "single", "unit": "tokens", "value": 40100 }, { "basis": null, "n": null, "name": "output_tokens", "scope": "per_person", "source_field": "team_load_requests[3].output_tokens", "statistic": "single", "unit": "tokens", "value": 512 }, { "basis": null, "n": null, "name": "needle_correct", "scope": "per_person", "source_field": "team_load_requests[3].needle_correct", "statistic": "single", "unit": "count", "value": 3 }, { "basis": null, "n": null, "name": "needle_total", "scope": "per_person", "source_field": "team_load_requests[3].needle_total", "statistic": "single", "unit": "count", "value": 3 }, { "basis": null, "n": null, "name": "ttft_s", "scope": "per_person", "source_field": "team_load_requests[4].ttft_s", "statistic": "single", "unit": "s", "value": 91.9306046019774 }, { "basis": null, "n": null, "name": "total_s", "scope": "per_person", "source_field": "team_load_requests[4].total_s", "statistic": "single", "unit": "s", "value": 199.8824033520068 }, { "basis": null, "n": null, "name": "decode_tps", "scope": "per_person", "source_field": "team_load_requests[4].decode_tps", "statistic": "single", "unit": "tok/s", "value": 4.733594121791887 }, { "basis": null, "n": null, "name": "user_number", "scope": "per_person", "source_field": "team_load_requests[4].user_number", "statistic": "single", "unit": "count", "value": 4 }, { "basis": null, "n": null, "name": "prompt_tokens", "scope": "per_person", "source_field": "team_load_requests[4].prompt_tokens", "statistic": "single", "unit": "tokens", "value": 40104 }, { "basis": null, "n": null, "name": "output_tokens", "scope": "per_person", "source_field": "team_load_requests[4].output_tokens", "statistic": "single", "unit": "tokens", "value": 512 }, { "basis": null, "n": null, "name": "needle_correct", "scope": "per_person", "source_field": "team_load_requests[4].needle_correct", "statistic": "single", "unit": "count", "value": 3 }, { "basis": null, "n": null, "name": "needle_total", "scope": "per_person", "source_field": "team_load_requests[4].needle_total", "statistic": "single", "unit": "count", "value": 3 }, { "basis": null, "n": null, "name": "ttft_s", "scope": "per_person", "source_field": "team_load_requests[5].ttft_s", "statistic": "single", "unit": "s", "value": 93.53399686899502 }, { "basis": null, "n": null, "name": "total_s", "scope": "per_person", "source_field": "team_load_requests[5].total_s", "statistic": "single", "unit": "s", "value": 199.94388840498868 }, { "basis": null, "n": null, "name": "decode_tps", "scope": "per_person", "source_field": "team_load_requests[5].decode_tps", "statistic": "single", "unit": "tok/s", "value": 4.802185141098013 }, { "basis": null, "n": null, "name": "user_number", "scope": "per_person", "source_field": "team_load_requests[5].user_number", "statistic": "single", "unit": "count", "value": 5 }, { "basis": null, "n": null, "name": "prompt_tokens", "scope": "per_person", "source_field": "team_load_requests[5].prompt_tokens", "statistic": "single", "unit": "tokens", "value": 40101 }, { "basis": null, "n": null, "name": "output_tokens", "scope": "per_person", "source_field": "team_load_requests[5].output_tokens", "statistic": "single", "unit": "tokens", "value": 512 }, { "basis": null, "n": null, "name": "needle_correct", "scope": "per_person", "source_field": "team_load_requests[5].needle_correct", "statistic": "single", "unit": "count", "value": 3 }, { "basis": null, "n": null, "name": "needle_total", "scope": "per_person", "source_field": "team_load_requests[5].needle_total", "statistic": "single", "unit": "count", "value": 3 }, { "basis": null, "n": null, "name": "ttft_s", "scope": "per_person", "source_field": "team_load_requests[6].ttft_s", "statistic": "single", "unit": "s", "value": 62.44563609204488 }, { "basis": null, "n": null, "name": "total_s", "scope": "per_person", "source_field": "team_load_requests[6].total_s", "statistic": "single", "unit": "s", "value": 198.7315115420497 }, { "basis": null, "n": null, "name": "decode_tps", "scope": "per_person", "source_field": "team_load_requests[6].decode_tps", "statistic": "single", "unit": "tok/s", "value": 3.7494714570583323 }, { "basis": null, "n": null, "name": "user_number", "scope": "per_person", "source_field": "team_load_requests[6].user_number", "statistic": "single", "unit": "count", "value": 6 }, { "basis": null, "n": null, "name": "prompt_tokens", "scope": "per_person", "source_field": "team_load_requests[6].prompt_tokens", "statistic": "single", "unit": "tokens", "value": 40106 }, { "basis": null, "n": null, "name": "output_tokens", "scope": "per_person", "source_field": "team_load_requests[6].output_tokens", "statistic": "single", "unit": "tokens", "value": 512 }, { "basis": null, "n": null, "name": "needle_correct", "scope": "per_person", "source_field": "team_load_requests[6].needle_correct", "statistic": "single", "unit": "count", "value": 3 }, { "basis": null, "n": null, "name": "needle_total", "scope": "per_person", "source_field": "team_load_requests[6].needle_total", "statistic": "single", "unit": "count", "value": 3 }, { "basis": null, "n": null, "name": "ttft_s", "scope": "per_person", "source_field": "team_load_requests[7].ttft_s", "statistic": "single", "unit": "s", "value": 60.890863093023654 }, { "basis": null, "n": null, "name": "total_s", "scope": "per_person", "source_field": "team_load_requests[7].total_s", "statistic": "single", "unit": "s", "value": 198.65202202403452 }, { "basis": null, "n": null, "name": "decode_tps", "scope": "per_person", "source_field": "team_load_requests[7].decode_tps", "statistic": "single", "unit": "tok/s", "value": 3.7093183881815532 }, { "basis": null, "n": null, "name": "user_number", "scope": "per_person", "source_field": "team_load_requests[7].user_number", "statistic": "single", "unit": "count", "value": 7 }, { "basis": null, "n": null, "name": "prompt_tokens", "scope": "per_person", "source_field": "team_load_requests[7].prompt_tokens", "statistic": "single", "unit": "tokens", "value": 40106 }, { "basis": null, "n": null, "name": "output_tokens", "scope": "per_person", "source_field": "team_load_requests[7].output_tokens", "statistic": "single", "unit": "tokens", "value": 512 }, { "basis": null, "n": null, "name": "needle_correct", "scope": "per_person", "source_field": "team_load_requests[7].needle_correct", "statistic": "single", "unit": "count", "value": 3 }, { "basis": null, "n": null, "name": "needle_total", "scope": "per_person", "source_field": "team_load_requests[7].needle_total", "statistic": "single", "unit": "count", "value": 3 }, { "basis": null, "n": null, "name": "ttft_s", "scope": "per_person", "source_field": "team_load_requests[8].ttft_s", "statistic": "single", "unit": "s", "value": 108.32101823401172 }, { "basis": null, "n": null, "name": "total_s", "scope": "per_person", "source_field": "team_load_requests[8].total_s", "statistic": "single", "unit": "s", "value": 200.42742160096532 }, { "basis": null, "n": null, "name": "decode_tps", "scope": "per_person", "source_field": "team_load_requests[8].decode_tps", "statistic": "single", "unit": "tok/s", "value": 5.547931319868898 }, { "basis": null, "n": null, "name": "user_number", "scope": "per_person", "source_field": "team_load_requests[8].user_number", "statistic": "single", "unit": "count", "value": 8 }, { "basis": null, "n": null, "name": "prompt_tokens", "scope": "per_person", "source_field": "team_load_requests[8].prompt_tokens", "statistic": "single", "unit": "tokens", "value": 40104 }, { "basis": null, "n": null, "name": "output_tokens", "scope": "per_person", "source_field": "team_load_requests[8].output_tokens", "statistic": "single", "unit": "tokens", "value": 512 }, { "basis": null, "n": null, "name": "needle_correct", "scope": "per_person", "source_field": "team_load_requests[8].needle_correct", "statistic": "single", "unit": "count", "value": 3 }, { "basis": null, "n": null, "name": "needle_total", "scope": "per_person", "source_field": "team_load_requests[8].needle_total", "statistic": "single", "unit": "count", "value": 3 }, { "basis": null, "n": null, "name": "ttft_s", "scope": "per_person", "source_field": "team_load_requests[9].ttft_s", "statistic": "single", "unit": "s", "value": 62.44542540400289 }, { "basis": null, "n": null, "name": "total_s", "scope": "per_person", "source_field": "team_load_requests[9].total_s", "statistic": "single", "unit": "s", "value": 198.73236480297055 }, { "basis": null, "n": null, "name": "decode_tps", "scope": "per_person", "source_field": "team_load_requests[9].decode_tps", "statistic": "single", "unit": "tok/s", "value": 3.749442186122427 }, { "basis": null, "n": null, "name": "user_number", "scope": "per_person", "source_field": "team_load_requests[9].user_number", "statistic": "single", "unit": "count", "value": 9 }, { "basis": null, "n": null, "name": "prompt_tokens", "scope": "per_person", "source_field": "team_load_requests[9].prompt_tokens", "statistic": "single", "unit": "tokens", "value": 40102 }, { "basis": null, "n": null, "name": "output_tokens", "scope": "per_person", "source_field": "team_load_requests[9].output_tokens", "statistic": "single", "unit": "tokens", "value": 512 }, { "basis": null, "n": null, "name": "needle_correct", "scope": "per_person", "source_field": "team_load_requests[9].needle_correct", "statistic": "single", "unit": "count", "value": 3 }, { "basis": null, "n": null, "name": "needle_total", "scope": "per_person", "source_field": "team_load_requests[9].needle_total", "statistic": "single", "unit": "count", "value": 3 }, { "basis": null, "n": null, "name": "ttft_s", "scope": "per_person", "source_field": "team_load_requests[10].ttft_s", "statistic": "single", "unit": "s", "value": 120.38016698602587 }, { "basis": null, "n": null, "name": "total_s", "scope": "per_person", "source_field": "team_load_requests[10].total_s", "statistic": "single", "unit": "s", "value": 200.78525526000885 }, { "basis": null, "n": null, "name": "decode_tps", "scope": "per_person", "source_field": "team_load_requests[10].decode_tps", "statistic": "single", "unit": "tok/s", "value": 6.355319184014209 }, { "basis": null, "n": null, "name": "user_number", "scope": "per_person", "source_field": "team_load_requests[10].user_number", "statistic": "single", "unit": "count", "value": 10 }, { "basis": null, "n": null, "name": "prompt_tokens", "scope": "per_person", "source_field": "team_load_requests[10].prompt_tokens", "statistic": "single", "unit": "tokens", "value": 40103 }, { "basis": null, "n": null, "name": "output_tokens", "scope": "per_person", "source_field": "team_load_requests[10].output_tokens", "statistic": "single", "unit": "tokens", "value": 512 }, { "basis": null, "n": null, "name": "needle_correct", "scope": "per_person", "source_field": "team_load_requests[10].needle_correct", "statistic": "single", "unit": "count", "value": 3 }, { "basis": null, "n": null, "name": "needle_total", "scope": "per_person", "source_field": "team_load_requests[10].needle_total", "statistic": "single", "unit": "count", "value": 3 }, { "basis": null, "n": null, "name": "ttft_s", "scope": "per_person", "source_field": "team_load_requests[11].ttft_s", "statistic": "single", "unit": "s", "value": 62.44689436000772 }, { "basis": null, "n": null, "name": "total_s", "scope": "per_person", "source_field": "team_load_requests[11].total_s", "statistic": "single", "unit": "s", "value": 198.72969399398426 }, { "basis": null, "n": null, "name": "decode_tps", "scope": "per_person", "source_field": "team_load_requests[11].decode_tps", "statistic": "single", "unit": "tok/s", "value": 3.7495560802421544 }, { "basis": null, "n": null, "name": "user_number", "scope": "per_person", "source_field": "team_load_requests[11].user_number", "statistic": "single", "unit": "count", "value": 11 }, { "basis": null, "n": null, "name": "prompt_tokens", "scope": "per_person", "source_field": "team_load_requests[11].prompt_tokens", "statistic": "single", "unit": "tokens", "value": 40099 }, { "basis": null, "n": null, "name": "output_tokens", "scope": "per_person", "source_field": "team_load_requests[11].output_tokens", "statistic": "single", "unit": "tokens", "value": 512 }, { "basis": null, "n": null, "name": "needle_correct", "scope": "per_person", "source_field": "team_load_requests[11].needle_correct", "statistic": "single", "unit": "count", "value": 3 }, { "basis": null, "n": null, "name": "needle_total", "scope": "per_person", "source_field": "team_load_requests[11].needle_total", "statistic": "single", "unit": "count", "value": 3 }, { "basis": null, "n": null, "name": "ttft_s", "scope": "per_person", "source_field": "team_load_requests[12].ttft_s", "statistic": "single", "unit": "s", "value": 162.46046369097894 }, { "basis": null, "n": null, "name": "total_s", "scope": "per_person", "source_field": "team_load_requests[12].total_s", "statistic": "single", "unit": "s", "value": 202.04564789996948 }, { "basis": null, "n": null, "name": "decode_tps", "scope": "per_person", "source_field": "team_load_requests[12].decode_tps", "statistic": "single", "unit": "tok/s", "value": 12.908870078819598 }, { "basis": null, "n": null, "name": "user_number", "scope": "per_person", "source_field": "team_load_requests[12].user_number", "statistic": "single", "unit": "count", "value": 12 }, { "basis": null, "n": null, "name": "prompt_tokens", "scope": "per_person", "source_field": "team_load_requests[12].prompt_tokens", "statistic": "single", "unit": "tokens", "value": 40102 }, { "basis": null, "n": null, "name": "output_tokens", "scope": "per_person", "source_field": "team_load_requests[12].output_tokens", "statistic": "single", "unit": "tokens", "value": 512 }, { "basis": null, "n": null, "name": "needle_correct", "scope": "per_person", "source_field": "team_load_requests[12].needle_correct", "statistic": "single", "unit": "count", "value": 3 }, { "basis": null, "n": null, "name": "needle_total", "scope": "per_person", "source_field": "team_load_requests[12].needle_total", "statistic": "single", "unit": "count", "value": 3 }, { "basis": null, "n": null, "name": "ttft_s", "scope": "per_person", "source_field": "team_load_requests[13].ttft_s", "statistic": "single", "unit": "s", "value": 146.2711280939984 }, { "basis": null, "n": null, "name": "total_s", "scope": "per_person", "source_field": "team_load_requests[13].total_s", "statistic": "single", "unit": "s", "value": 201.5089382849983 }, { "basis": null, "n": null, "name": "decode_tps", "scope": "per_person", "source_field": "team_load_requests[13].decode_tps", "statistic": "single", "unit": "tok/s", "value": 9.250909806762385 }, { "basis": null, "n": null, "name": "user_number", "scope": "per_person", "source_field": "team_load_requests[13].user_number", "statistic": "single", "unit": "count", "value": 13 }, { "basis": null, "n": null, "name": "prompt_tokens", "scope": "per_person", "source_field": "team_load_requests[13].prompt_tokens", "statistic": "single", "unit": "tokens", "value": 40099 }, { "basis": null, "n": null, "name": "output_tokens", "scope": "per_person", "source_field": "team_load_requests[13].output_tokens", "statistic": "single", "unit": "tokens", "value": 512 }, { "basis": null, "n": null, "name": "needle_correct", "scope": "per_person", "source_field": "team_load_requests[13].needle_correct", "statistic": "single", "unit": "count", "value": 3 }, { "basis": null, "n": null, "name": "needle_total", "scope": "per_person", "source_field": "team_load_requests[13].needle_total", "statistic": "single", "unit": "count", "value": 3 }, { "basis": null, "n": null, "name": "ttft_s", "scope": "per_person", "source_field": "team_load_requests[14].ttft_s", "statistic": "single", "unit": "s", "value": 126.69506250100676 }, { "basis": null, "n": null, "name": "total_s", "scope": "per_person", "source_field": "team_load_requests[14].total_s", "statistic": "single", "unit": "s", "value": 200.97121076699113 }, { "basis": null, "n": null, "name": "decode_tps", "scope": "per_person", "source_field": "team_load_requests[14].decode_tps", "statistic": "single", "unit": "tok/s", "value": 6.879732079941716 }, { "basis": null, "n": null, "name": "user_number", "scope": "per_person", "source_field": "team_load_requests[14].user_number", "statistic": "single", "unit": "count", "value": 14 }, { "basis": null, "n": null, "name": "prompt_tokens", "scope": "per_person", "source_field": "team_load_requests[14].prompt_tokens", "statistic": "single", "unit": "tokens", "value": 40100 }, { "basis": null, "n": null, "name": "output_tokens", "scope": "per_person", "source_field": "team_load_requests[14].output_tokens", "statistic": "single", "unit": "tokens", "value": 512 }, { "basis": null, "n": null, "name": "needle_correct", "scope": "per_person", "source_field": "team_load_requests[14].needle_correct", "statistic": "single", "unit": "count", "value": 3 }, { "basis": null, "n": null, "name": "needle_total", "scope": "per_person", "source_field": "team_load_requests[14].needle_total", "statistic": "single", "unit": "count", "value": 3 }, { "basis": null, "n": null, "name": "ttft_s", "scope": "per_person", "source_field": "team_load_requests[15].ttft_s", "statistic": "single", "unit": "s", "value": 138.4330267990008 }, { "basis": null, "n": null, "name": "total_s", "scope": "per_person", "source_field": "team_load_requests[15].total_s", "statistic": "single", "unit": "s", "value": 201.29494035604876 }, { "basis": null, "n": null, "name": "decode_tps", "scope": "per_person", "source_field": "team_load_requests[15].decode_tps", "statistic": "single", "unit": "tok/s", "value": 8.128928489207718 }, { "basis": null, "n": null, "name": "user_number", "scope": "per_person", "source_field": "team_load_requests[15].user_number", "statistic": "single", "unit": "count", "value": 15 }, { "basis": null, "n": null, "name": "prompt_tokens", "scope": "per_person", "source_field": "team_load_requests[15].prompt_tokens", "statistic": "single", "unit": "tokens", "value": 40104 }, { "basis": null, "n": null, "name": "output_tokens", "scope": "per_person", "source_field": "team_load_requests[15].output_tokens", "statistic": "single", "unit": "tokens", "value": 512 }, { "basis": null, "n": null, "name": "needle_correct", "scope": "per_person", "source_field": "team_load_requests[15].needle_correct", "statistic": "single", "unit": "count", "value": 3 }, { "basis": null, "n": null, "name": "needle_total", "scope": "per_person", "source_field": "team_load_requests[15].needle_total", "statistic": "single", "unit": "count", "value": 3 }, { "basis": null, "n": null, "name": "ttft_s", "scope": "per_person", "source_field": "team_load_requests[16].ttft_s", "statistic": "single", "unit": "s", "value": 159.39992942800745 }, { "basis": null, "n": null, "name": "total_s", "scope": "per_person", "source_field": "team_load_requests[16].total_s", "statistic": "single", "unit": "s", "value": 201.91818897501798 }, { "basis": null, "n": null, "name": "decode_tps", "scope": "per_person", "source_field": "team_load_requests[16].decode_tps", "statistic": "single", "unit": "tok/s", "value": 12.018365884309311 }, { "basis": null, "n": null, "name": "user_number", "scope": "per_person", "source_field": "team_load_requests[16].user_number", "statistic": "single", "unit": "count", "value": 16 }, { "basis": null, "n": null, "name": "prompt_tokens", "scope": "per_person", "source_field": "team_load_requests[16].prompt_tokens", "statistic": "single", "unit": "tokens", "value": 40105 }, { "basis": null, "n": null, "name": "output_tokens", "scope": "per_person", "source_field": "team_load_requests[16].output_tokens", "statistic": "single", "unit": "tokens", "value": 512 }, { "basis": null, "n": null, "name": "needle_correct", "scope": "per_person", "source_field": "team_load_requests[16].needle_correct", "statistic": "single", "unit": "count", "value": 3 }, { "basis": null, "n": null, "name": "needle_total", "scope": "per_person", "source_field": "team_load_requests[16].needle_total", "statistic": "single", "unit": "count", "value": 3 }, { "basis": null, "n": null, "name": "ttft_s", "scope": "per_person", "source_field": "team_load_requests[17].ttft_s", "statistic": "single", "unit": "s", "value": 149.4206585349748 }, { "basis": null, "n": null, "name": "total_s", "scope": "per_person", "source_field": "team_load_requests[17].total_s", "statistic": "single", "unit": "s", "value": 201.60680380096892 }, { "basis": null, "n": null, "name": "decode_tps", "scope": "per_person", "source_field": "team_load_requests[17].decode_tps", "statistic": "single", "unit": "tok/s", "value": 9.791870953399984 }, { "basis": null, "n": null, "name": "user_number", "scope": "per_person", "source_field": "team_load_requests[17].user_number", "statistic": "single", "unit": "count", "value": 17 }, { "basis": null, "n": null, "name": "prompt_tokens", "scope": "per_person", "source_field": "team_load_requests[17].prompt_tokens", "statistic": "single", "unit": "tokens", "value": 40102 }, { "basis": null, "n": null, "name": "output_tokens", "scope": "per_person", "source_field": "team_load_requests[17].output_tokens", "statistic": "single", "unit": "tokens", "value": 512 }, { "basis": null, "n": null, "name": "needle_correct", "scope": "per_person", "source_field": "team_load_requests[17].needle_correct", "statistic": "single", "unit": "count", "value": 3 }, { "basis": null, "n": null, "name": "needle_total", "scope": "per_person", "source_field": "team_load_requests[17].needle_total", "statistic": "single", "unit": "count", "value": 3 }, { "basis": null, "n": null, "name": "ttft_s", "scope": "per_person", "source_field": "team_load_requests[18].ttft_s", "statistic": "single", "unit": "s", "value": 59.14211418799823 }, { "basis": null, "n": null, "name": "total_s", "scope": "per_person", "source_field": "team_load_requests[18].total_s", "statistic": "single", "unit": "s", "value": 198.56782370398287 }, { "basis": null, "n": null, "name": "decode_tps", "scope": "per_person", "source_field": "team_load_requests[18].decode_tps", "statistic": "single", "unit": "tok/s", "value": 3.66503424493182 }, { "basis": null, "n": null, "name": "user_number", "scope": "per_person", "source_field": "team_load_requests[18].user_number", "statistic": "single", "unit": "count", "value": 18 }, { "basis": null, "n": null, "name": "prompt_tokens", "scope": "per_person", "source_field": "team_load_requests[18].prompt_tokens", "statistic": "single", "unit": "tokens", "value": 40099 }, { "basis": null, "n": null, "name": "output_tokens", "scope": "per_person", "source_field": "team_load_requests[18].output_tokens", "statistic": "single", "unit": "tokens", "value": 512 }, { "basis": null, "n": null, "name": "needle_correct", "scope": "per_person", "source_field": "team_load_requests[18].needle_correct", "statistic": "single", "unit": "count", "value": 3 }, { "basis": null, "n": null, "name": "needle_total", "scope": "per_person", "source_field": "team_load_requests[18].needle_total", "statistic": "single", "unit": "count", "value": 3 }, { "basis": null, "n": null, "name": "ttft_s", "scope": "per_person", "source_field": "team_load_requests[19].ttft_s", "statistic": "single", "unit": "s", "value": 62.44722560798982 }, { "basis": null, "n": null, "name": "total_s", "scope": "per_person", "source_field": "team_load_requests[19].total_s", "statistic": "single", "unit": "s", "value": 198.7314917339827 }, { "basis": null, "n": null, "name": "decode_tps", "scope": "per_person", "source_field": "team_load_requests[19].decode_tps", "statistic": "single", "unit": "tok/s", "value": 3.7495157330016933 }, { "basis": null, "n": null, "name": "user_number", "scope": "per_person", "source_field": "team_load_requests[19].user_number", "statistic": "single", "unit": "count", "value": 19 }, { "basis": null, "n": null, "name": "prompt_tokens", "scope": "per_person", "source_field": "team_load_requests[19].prompt_tokens", "statistic": "single", "unit": "tokens", "value": 40101 }, { "basis": null, "n": null, "name": "output_tokens", "scope": "per_person", "source_field": "team_load_requests[19].output_tokens", "statistic": "single", "unit": "tokens", "value": 512 }, { "basis": null, "n": null, "name": "needle_correct", "scope": "per_person", "source_field": "team_load_requests[19].needle_correct", "statistic": "single", "unit": "count", "value": 3 }, { "basis": null, "n": null, "name": "needle_total", "scope": "per_person", "source_field": "team_load_requests[19].needle_total", "statistic": "single", "unit": "count", "value": 3 }, { "basis": null, "n": 20, "name": "decode_tps", "scope": "per_person", "source_field": "per_person_decode_median_tps", "statistic": "median", "unit": "tok/s", "value": 5.951625251941554 } ], "model": { "artifact_sha256": [ "07141c2db92e47bc08777132cd1a0323faf300eab3a7d7c111bc2bf075fda050", "6dea9c759a0f941cf9e1cc1501216b0e107966a66c90425050e212efbd053f02", "9758875fc55e49561165f4a44342b654c123c3e25c6811b34abb83553fb1a164" ], "publisher_repo": "Qwen/Qwen3.6-35B-A3B", "publisher_revision": "995ad96eacd98c81ed38be0c5b274b04031597b0", "variant": "NVFP4 variant", "weights_dtype": "nvfp4", "weights_format": "safetensors" }, "outcome": { "gates": [ { "met": true, "name": "requests", "observed": "20/20" }, { "met": true, "name": "hidden-facts", "observed": "60/60" }, { "met": true, "name": "correctness-probes", "observed": "20/20" } ], "reasons": [], "status": "pass" }, "publication_group": "pre-standard", "published": null, "publisher": { "name": "Skilak Keep" }, "quality": [ { "passed": 60, "skipped": null, "suite": "hidden-facts", "total": 60, "version": "1" }, { "passed": 20, "skipped": null, "suite": "correctness-probes", "total": 20, "version": "1" } ], "record_date": "2026-10-07", "reproduce": { "driver": null, "fixture_sha256": "e2097959b1ca0142f2b93edc3ec411e474c0a6622800dd247af7a71bd9c39e83", "harness": { "name": "keep-team-load-v1", "version": "1" }, "harness_sha256": null, "os": null, "workload_name": null, "workload_version": null }, "requests": [ { "cached_prompt_tokens": null, "checks_passed": true, "completed": true, "decode_tps": 4.802075908989307, "effective_prefill_tps": null, "needle_correct": 3, "needle_total": 3, "output_tokens": 512, "prompt_tokens": 40100, "total_s": 199.94406931602862, "ttft_s": 93.53175729000941, "user_number": 0 }, { "cached_prompt_tokens": null, "checks_passed": true, "completed": true, "decode_tps": 12.810027045889365, "effective_prefill_tps": null, "needle_correct": 3, "needle_total": 3, "output_tokens": 512, "prompt_tokens": 40103, "total_s": 202.01941937801894, "ttft_s": 162.12879321700893, "user_number": 1 }, { "cached_prompt_tokens": null, "checks_passed": true, "completed": true, "decode_tps": 10.72207718609505, "effective_prefill_tps": null, "needle_correct": 3, "needle_total": 3, "output_tokens": 512, "prompt_tokens": 40102, "total_s": 201.7493742229999, "ttft_s": 154.09069847100182, "user_number": 2 }, { "cached_prompt_tokens": null, "checks_passed": true, "completed": true, "decode_tps": 6.746349765155457, "effective_prefill_tps": null, "needle_correct": 3, "needle_total": 3, "output_tokens": 512, "prompt_tokens": 40100, "total_s": 200.92618343001232, "ttft_s": 125.18151886499254, "user_number": 3 }, { "cached_prompt_tokens": null, "checks_passed": true, "completed": true, "decode_tps": 4.733594121791887, "effective_prefill_tps": null, "needle_correct": 3, "needle_total": 3, "output_tokens": 512, "prompt_tokens": 40104, "total_s": 199.8824033520068, "ttft_s": 91.9306046019774, "user_number": 4 }, { "cached_prompt_tokens": null, "checks_passed": true, "completed": true, "decode_tps": 4.802185141098013, "effective_prefill_tps": null, "needle_correct": 3, "needle_total": 3, "output_tokens": 512, "prompt_tokens": 40101, "total_s": 199.94388840498868, "ttft_s": 93.53399686899502, "user_number": 5 }, { "cached_prompt_tokens": null, "checks_passed": true, "completed": true, "decode_tps": 3.7494714570583323, "effective_prefill_tps": null, "needle_correct": 3, "needle_total": 3, "output_tokens": 512, "prompt_tokens": 40106, "total_s": 198.7315115420497, "ttft_s": 62.44563609204488, "user_number": 6 }, { "cached_prompt_tokens": null, "checks_passed": true, "completed": true, "decode_tps": 3.7093183881815532, "effective_prefill_tps": null, "needle_correct": 3, "needle_total": 3, "output_tokens": 512, "prompt_tokens": 40106, "total_s": 198.65202202403452, "ttft_s": 60.890863093023654, "user_number": 7 }, { "cached_prompt_tokens": null, "checks_passed": true, "completed": true, "decode_tps": 5.547931319868898, "effective_prefill_tps": null, "needle_correct": 3, "needle_total": 3, "output_tokens": 512, "prompt_tokens": 40104, "total_s": 200.42742160096532, "ttft_s": 108.32101823401172, "user_number": 8 }, { "cached_prompt_tokens": null, "checks_passed": true, "completed": true, "decode_tps": 3.749442186122427, "effective_prefill_tps": null, "needle_correct": 3, "needle_total": 3, "output_tokens": 512, "prompt_tokens": 40102, "total_s": 198.73236480297055, "ttft_s": 62.44542540400289, "user_number": 9 }, { "cached_prompt_tokens": null, "checks_passed": true, "completed": true, "decode_tps": 6.355319184014209, "effective_prefill_tps": null, "needle_correct": 3, "needle_total": 3, "output_tokens": 512, "prompt_tokens": 40103, "total_s": 200.78525526000885, "ttft_s": 120.38016698602587, "user_number": 10 }, { "cached_prompt_tokens": null, "checks_passed": true, "completed": true, "decode_tps": 3.7495560802421544, "effective_prefill_tps": null, "needle_correct": 3, "needle_total": 3, "output_tokens": 512, "prompt_tokens": 40099, "total_s": 198.72969399398426, "ttft_s": 62.44689436000772, "user_number": 11 }, { "cached_prompt_tokens": null, "checks_passed": true, "completed": true, "decode_tps": 12.908870078819598, "effective_prefill_tps": null, "needle_correct": 3, "needle_total": 3, "output_tokens": 512, "prompt_tokens": 40102, "total_s": 202.04564789996948, "ttft_s": 162.46046369097894, "user_number": 12 }, { "cached_prompt_tokens": null, "checks_passed": true, "completed": true, "decode_tps": 9.250909806762385, "effective_prefill_tps": null, "needle_correct": 3, "needle_total": 3, "output_tokens": 512, "prompt_tokens": 40099, "total_s": 201.5089382849983, "ttft_s": 146.2711280939984, "user_number": 13 }, { "cached_prompt_tokens": null, "checks_passed": true, "completed": true, "decode_tps": 6.879732079941716, "effective_prefill_tps": null, "needle_correct": 3, "needle_total": 3, "output_tokens": 512, "prompt_tokens": 40100, "total_s": 200.97121076699113, "ttft_s": 126.69506250100676, "user_number": 14 }, { "cached_prompt_tokens": null, "checks_passed": true, "completed": true, "decode_tps": 8.128928489207718, "effective_prefill_tps": null, "needle_correct": 3, "needle_total": 3, "output_tokens": 512, "prompt_tokens": 40104, "total_s": 201.29494035604876, "ttft_s": 138.4330267990008, "user_number": 15 }, { "cached_prompt_tokens": null, "checks_passed": true, "completed": true, "decode_tps": 12.018365884309311, "effective_prefill_tps": null, "needle_correct": 3, "needle_total": 3, "output_tokens": 512, "prompt_tokens": 40105, "total_s": 201.91818897501798, "ttft_s": 159.39992942800745, "user_number": 16 }, { "cached_prompt_tokens": null, "checks_passed": true, "completed": true, "decode_tps": 9.791870953399984, "effective_prefill_tps": null, "needle_correct": 3, "needle_total": 3, "output_tokens": 512, "prompt_tokens": 40102, "total_s": 201.60680380096892, "ttft_s": 149.4206585349748, "user_number": 17 }, { "cached_prompt_tokens": null, "checks_passed": true, "completed": true, "decode_tps": 3.66503424493182, "effective_prefill_tps": null, "needle_correct": 3, "needle_total": 3, "output_tokens": 512, "prompt_tokens": 40099, "total_s": 198.56782370398287, "ttft_s": 59.14211418799823, "user_number": 18 }, { "cached_prompt_tokens": null, "checks_passed": true, "completed": true, "decode_tps": 3.7495157330016933, "effective_prefill_tps": null, "needle_correct": 3, "needle_total": 3, "output_tokens": 512, "prompt_tokens": 40101, "total_s": 198.7314917339827, "ttft_s": 62.44722560798982, "user_number": 19 } ], "scenario": { "id": "team-load", "version": "1" }, "settings": { "concurrency": 20, "configured_context_tokens": 262144, "context_tokens": 262144, "kv_dtype": "fp8", "max_output_tokens": 512, "speculative_decoding": "plain", "thinking": false }, "setup_complete": false, "setup_id": "a4b96c03f67c006a12a64bbc7734ee71", "standard": "pre-standard", "system": { "driver": null, "hardware": "NVIDIA GB10, 128 GB unified memory", "os": null }, "workload": { "cache": null, "content_tokens": null, "document": null, "input_tokens": { "max": 40106, "median": 40102, "min": 40099 }, "output_tokens": 512, "repetitions": null, "users": 20, "warmup": null } }