{
  "base_url": "http://127.0.0.1:39118/v1",
  "completed": 60,
  "concurrency": 1,
  "context_distribution": {
    "4096": 60
  },
  "context_seed": 0,
  "context_tokens": 4096,
  "engine": "sglang-glm53-sm120-v0.1.1-rc14",
  "failed": 0,
  "failures": [],
  "finished_at": "2026-09-02T11:40:41Z",
  "gpu": "2x RTX PRO 6000 Blackwell Max-Q TP2 WSL2",
  "manifest": null,
  "max_context_tokens": 245760,
  "max_tokens": 256,
  "measurement_protocol": "capacity-v3",
  "metrics": {
    "content_chunks": 1453,
    "content_chunks_s": null,
    "decode_tok_s_p50": 105.82483220207068,
    "decode_tok_s_p95": 123.67263070012335,
    "e2e_p50_ms": 602.8908000444062,
    "e2e_p95_ms": 1636.0194999724627,
    "effective_prefill_tok_s_p50": 16968.594099022444,
    "effective_prefill_tok_s_p95": 17566.92471532026,
    "generation_p50_ms": 421.8983000027947,
    "generation_p95_ms": 1459.9709999747574,
    "mean_inter_token_latency_ms_p50": 9.423787499794603,
    "mean_inter_token_latency_ms_p95": 12.22221627912679,
    "output_token_sources": {
      "usage": 60
    },
    "output_tokens": 3560,
    "prefix_cache_hit_avg": 0.9915796564499831,
    "prompt_token_samples": 60,
    "prompt_tokens": 178140,
    "prompt_tokens_p50": 2969.0,
    "prompt_tokens_p95": 2969.0,
    "reasoning_chunks": 0,
    "throughput_tok_s": 81.39120877457472,
    "time_to_first_output_p50_ms": 174.9363000271842,
    "time_to_first_output_p95_ms": 183.6949000135064,
    "ttft_p50_ms": 174.93670003022999,
    "ttft_p95_ms": 183.6953000165522,
    "ttft_samples": 60
  },
  "model": "glm53-flash-ormandj-sglang-sm120-tp2-240k-c1-adaptive-mtp",
  "request_timings": [
    {
      "content_chunks": 110,
      "decode_tok_s": 97.31338014599827,
      "e2e_ms": 3092.2373999492265,
      "effective_prefill_tok_s": 5167.090731462181,
      "generation_ms": 2517.639399971813,
      "mean_inter_token_latency_ms": 10.276079183558421,
      "output_token_source": "usage",
      "output_tokens": 246,
      "planned_context_tokens": 4096,
      "prompt_tokens": 2969,
      "reasoning_chunks": 0,
      "request_index": 0,
      "time_to_first_output_ms": 574.5979999774136,
      "ttft_ms": 574.5987999835052,
      "visible_generation_ms": 2517.6385999657214
    },
    {
      "content_chunks": 15,
      "decode_tok_s": 115.07323047096239,
      "e2e_ms": 508.5294999880716,
      "effective_prefill_tok_s": 16651.243660171443,
      "generation_ms": 330.22450003772974,
      "mean_inter_token_latency_ms": 8.69011842204552,
      "output_token_source": "usage",
      "output_tokens": 39,
      "planned_context_tokens": 4096,
      "prompt_tokens": 2969,
      "reasoning_chunks": 0,
      "request_index": 1,
      "time_to_first_output_ms": 178.30499995034188,
      "ttft_ms": 178.30549995414913,
      "visible_generation_ms": 330.2240000339225
    },
    {
      "content_chunks": 16,
      "decode_tok_s": 114.00405327770302,
      "e2e_ms": 519.1925000399351,
      "effective_prefill_tok_s": 16764.59660267289,
      "generation_ms": 342.09310001460835,
      "mean_inter_token_latency_ms": 8.771617949092521,
      "output_token_source": "usage",
      "output_tokens": 40,
      "planned_context_tokens": 4096,
      "prompt_tokens": 2969,
      "reasoning_chunks": 0,
      "request_index": 2,
      "time_to_first_output_ms": 177.09940002532676,
      "ttft_ms": 177.0997000276111,
      "visible_generation_ms": 342.092800012324
    },
    {
      "content_chunks": 16,
      "decode_tok_s": 111.09474759788327,
      "e2e_ms": 527.834799955599,
      "effective_prefill_tok_s": 16794.591790410348,
      "generation_ms": 351.0516999522224,
      "mean_inter_token_latency_ms": 9.001325639800575,
      "output_token_source": "usage",
      "output_tokens": 40,
      "planned_context_tokens": 4096,
      "prompt_tokens": 2969,
      "reasoning_chunks": 0,
      "request_index": 3,
      "time_to_first_output_ms": 176.7831000033766,
      "ttft_ms": 176.78340000566095,
      "visible_generation_ms": 351.05139994993806
    },
    {
      "content_chunks": 17,
      "decode_tok_s": 103.07589315784017,
      "e2e_ms": 538.7595999636687,
      "effective_prefill_tok_s": 17454.52065799191,
      "generation_ms": 368.66039998130873,
      "mean_inter_token_latency_ms": 9.701589473192335,
      "output_token_source": "usage",
      "output_tokens": 39,
      "planned_context_tokens": 4096,
      "prompt_tokens": 2969,
      "reasoning_chunks": 0,
      "request_index": 4,
      "time_to_first_output_ms": 170.09919998236,
      "ttft_ms": 170.09949998464435,
      "visible_generation_ms": 368.6600999790244
    },
    {
      "content_chunks": 23,
      "decode_tok_s": 110.18259460082807,
      "e2e_ms": 675.7612000219524,
      "effective_prefill_tok_s": 16812.975603603347,
      "generation_ms": 499.1713999770582,
      "mean_inter_token_latency_ms": 9.075843635946512,
      "output_token_source": "usage",
      "output_tokens": 56,
      "planned_context_tokens": 4096,
      "prompt_tokens": 2969,
      "reasoning_chunks": 0,
      "request_index": 5,
      "time_to_first_output_ms": 176.58980004489422,
      "ttft_ms": 176.59020004794002,
      "visible_generation_ms": 499.1709999740124
    },
    {
      "content_chunks": 16,
      "decode_tok_s": 110.04480515963036,
      "e2e_ms": 527.342299988959,
      "effective_prefill_tok_s": 17167.684742118185,
      "generation_ms": 354.4011000194587,
      "mean_inter_token_latency_ms": 9.087207692806633,
      "output_token_source": "usage",
      "output_tokens": 40,
      "planned_context_tokens": 4096,
      "prompt_tokens": 2969,
      "reasoning_chunks": 0,
      "request_index": 6,
      "time_to_first_output_ms": 172.9411999695003,
      "ttft_ms": 172.9413999710232,
      "visible_generation_ms": 354.4009000179358
    },
    {
      "content_chunks": 64,
      "decode_tok_s": 119.18044947674196,
      "e2e_ms": 1636.0194999724627,
      "effective_prefill_tok_s": 16864.67081536452,
      "generation_ms": 1459.9709999747574,
      "mean_inter_token_latency_ms": 8.39063793088941,
      "output_token_source": "usage",
      "output_tokens": 175,
      "planned_context_tokens": 4096,
      "prompt_tokens": 2969,
      "reasoning_chunks": 0,
      "request_index": 7,
      "time_to_first_output_ms": 176.04849999770522,
      "ttft_ms": 176.04890000075102,
      "visible_generation_ms": 1459.9705999717116
    },
    {
      "content_chunks": 16,
      "decode_tok_s": 114.21348784779103,
      "e2e_ms": 602.8908000444062,
      "effective_prefill_tok_s": 11356.98574972303,
      "generation_ms": 341.46580001106486,
      "mean_inter_token_latency_ms": 8.755533333617048,
      "output_token_source": "usage",
      "output_tokens": 40,
      "planned_context_tokens": 4096,
      "prompt_tokens": 2969,
      "reasoning_chunks": 0,
      "request_index": 8,
      "time_to_first_output_ms": 261.42500003334135,
      "ttft_ms": 261.4253000356257,
      "visible_generation_ms": 341.4655000087805
    },
    {
      "content_chunks": 20,
      "decode_tok_s": 100.48264436235624,
      "e2e_ms": 631.6596999531612,
      "effective_prefill_tok_s": 17076.054874158086,
      "generation_ms": 457.790499960538,
      "mean_inter_token_latency_ms": 9.951967390446478,
      "output_token_source": "usage",
      "output_tokens": 47,
      "planned_context_tokens": 4096,
      "prompt_tokens": 2969,
      "reasoning_chunks": 0,
      "request_index": 9,
      "time_to_first_output_ms": 173.86919999262318,
      "ttft_ms": 173.86949999490753,
      "visible_generation_ms": 457.79019995825365
    },
    {
      "content_chunks": 19,
      "decode_tok_s": 105.00655263310775,
      "e2e_ms": 606.8622000166215,
      "effective_prefill_tok_s": 17589.45651603737,
      "generation_ms": 438.0679000169039,
      "mean_inter_token_latency_ms": 9.52321521775878,
      "output_token_source": "usage",
      "output_tokens": 47,
      "planned_context_tokens": 4096,
      "prompt_tokens": 2969,
      "reasoning_chunks": 0,
      "request_index": 10,
      "time_to_first_output_ms": 168.79429999971762,
      "ttft_ms": 168.79460000200197,
      "visible_generation_ms": 438.06760001461953
    },
    {
      "content_chunks": 18,
      "decode_tok_s": 97.0484650356555,
      "e2e_ms": 585.8915999997407,
      "effective_prefill_tok_s": 17090.091084649568,
      "generation_ms": 412.16519998852164,
      "mean_inter_token_latency_ms": 10.304129999713041,
      "output_token_source": "usage",
      "output_tokens": 41,
      "planned_context_tokens": 4096,
      "prompt_tokens": 2969,
      "reasoning_chunks": 0,
      "request_index": 11,
      "time_to_first_output_ms": 173.72640001121908,
      "ttft_ms": 173.72670001350343,
      "visible_generation_ms": 412.1648999862373
    },
    {
      "content_chunks": 27,
      "decode_tok_s": 130.20876806121953,
      "e2e_ms": 774.9330999795347,
      "effective_prefill_tok_s": 16879.378675726235,
      "generation_ms": 599.0379999857396,
      "mean_inter_token_latency_ms": 7.679974358791533,
      "output_token_source": "usage",
      "output_tokens": 79,
      "planned_context_tokens": 4096,
      "prompt_tokens": 2969,
      "reasoning_chunks": 0,
      "request_index": 12,
      "time_to_first_output_ms": 175.89509999379516,
      "ttft_ms": 175.8953999960795,
      "visible_generation_ms": 599.0376999834552
    },
    {
      "content_chunks": 15,
      "decode_tok_s": 122.67939345595832,
      "e2e_ms": 494.96400001225993,
      "effective_prefill_tok_s": 16768.11877094533,
      "generation_ms": 317.901799979154,
      "mean_inter_token_latency_ms": 8.151328204593693,
      "output_token_source": "usage",
      "output_tokens": 40,
      "planned_context_tokens": 4096,
      "prompt_tokens": 2969,
      "reasoning_chunks": 0,
      "request_index": 13,
      "time_to_first_output_ms": 177.0622000331059,
      "ttft_ms": 177.06250003539026,
      "visible_generation_ms": 317.9014999768697
    },
    {
      "content_chunks": 18,
      "decode_tok_s": 82.95838120174496,
      "e2e_ms": 590.6622000038624,
      "effective_prefill_tok_s": 17592.624962928785,
      "generation_ms": 421.8983000027947,
      "mean_inter_token_latency_ms": 12.054237142936993,
      "output_token_source": "usage",
      "output_tokens": 36,
      "planned_context_tokens": 4096,
      "prompt_tokens": 2969,
      "reasoning_chunks": 0,
      "request_index": 14,
      "time_to_first_output_ms": 168.76390000106767,
      "ttft_ms": 168.76420000335202,
      "visible_generation_ms": 421.89800000051036
    },
    {
      "content_chunks": 14,
      "decode_tok_s": 123.67263070012335,
      "e2e_ms": 508.1357000162825,
      "effective_prefill_tok_s": 16810.54812292716,
      "generation_ms": 331.52040001004934,
      "mean_inter_token_latency_ms": 8.085863414879253,
      "output_token_source": "usage",
      "output_tokens": 42,
      "planned_context_tokens": 4096,
      "prompt_tokens": 2969,
      "reasoning_chunks": 0,
      "request_index": 15,
      "time_to_first_output_ms": 176.61530000623316,
      "ttft_ms": 176.61590001080185,
      "visible_generation_ms": 331.51980000548065
    },
    {
      "content_chunks": 16,
      "decode_tok_s": 90.55329393678095,
      "e2e_ms": 552.9780000215396,
      "effective_prefill_tok_s": 16725.95960219817,
      "generation_ms": 375.4695000243373,
      "mean_inter_token_latency_ms": 11.043220588951097,
      "output_token_source": "usage",
      "output_tokens": 35,
      "planned_context_tokens": 4096,
      "prompt_tokens": 2969,
      "reasoning_chunks": 0,
      "request_index": 16,
      "time_to_first_output_ms": 177.5084999972023,
      "ttft_ms": 177.50879999948665,
      "visible_generation_ms": 375.469200022053
    },
    {
      "content_chunks": 23,
      "decode_tok_s": 121.26180662348972,
      "e2e_ms": 679.7550999908708,
      "effective_prefill_tok_s": 16801.415644201636,
      "generation_ms": 503.0438000103459,
      "mean_inter_token_latency_ms": 8.246619672300753,
      "output_token_source": "usage",
      "output_tokens": 62,
      "planned_context_tokens": 4096,
      "prompt_tokens": 2969,
      "reasoning_chunks": 0,
      "request_index": 17,
      "time_to_first_output_ms": 176.71129998052493,
      "ttft_ms": 176.71169998357072,
      "visible_generation_ms": 503.0434000073001
    },
    {
      "content_chunks": 18,
      "decode_tok_s": 92.81917539194882,
      "e2e_ms": 598.1979999924079,
      "effective_prefill_tok_s": 16677.320530995577,
      "generation_ms": 420.1718000113033,
      "mean_inter_token_latency_ms": 10.773635897725725,
      "output_token_source": "usage",
      "output_tokens": 40,
      "planned_context_tokens": 4096,
      "prompt_tokens": 2969,
      "reasoning_chunks": 0,
      "request_index": 18,
      "time_to_first_output_ms": 178.02619998110458,
      "ttft_ms": 178.02659998415038,
      "visible_generation_ms": 420.1714000082575
    },
    {
      "content_chunks": 32,
      "decode_tok_s": 111.24599883686382,
      "e2e_ms": 888.779099972453,
      "effective_prefill_tok_s": 16619.90627494852,
      "generation_ms": 710.1379000232555,
      "mean_inter_token_latency_ms": 8.989087342066526,
      "output_token_source": "usage",
      "output_tokens": 80,
      "planned_context_tokens": 4096,
      "prompt_tokens": 2969,
      "reasoning_chunks": 0,
      "request_index": 19,
      "time_to_first_output_ms": 178.64119994919747,
      "ttft_ms": 178.64169995300472,
      "visible_generation_ms": 710.1374000194483
    },
    {
      "content_chunks": 19,
      "decode_tok_s": 102.5808895888586,
      "e2e_ms": 617.01320001157,
      "effective_prefill_tok_s": 16648.44253753244,
      "generation_ms": 438.6782000074163,
      "mean_inter_token_latency_ms": 9.748404444609251,
      "output_token_source": "usage",
      "output_tokens": 46,
      "planned_context_tokens": 4096,
      "prompt_tokens": 2969,
      "reasoning_chunks": 0,
      "request_index": 20,
      "time_to_first_output_ms": 178.3350000041537,
      "ttft_ms": 178.33550000796095,
      "visible_generation_ms": 438.67770000360906
    },
    {
      "content_chunks": 19,
      "decode_tok_s": 98.86823462760204,
      "e2e_ms": 615.76029995922,
      "effective_prefill_tok_s": 16418.009491778208,
      "generation_ms": 434.92229998810217,
      "mean_inter_token_latency_ms": 10.114472092746562,
      "output_token_source": "usage",
      "output_tokens": 44,
      "planned_context_tokens": 4096,
      "prompt_tokens": 2969,
      "reasoning_chunks": 0,
      "request_index": 21,
      "time_to_first_output_ms": 180.83799997111782,
      "ttft_ms": 180.83859997568652,
      "visible_generation_ms": 434.9216999835335
    },
    {
      "content_chunks": 85,
      "decode_tok_s": 105.82483220207068,
      "e2e_ms": 2097.966799978167,
      "effective_prefill_tok_s": 17438.763158804326,
      "generation_ms": 1927.7138999896124,
      "mean_inter_token_latency_ms": 9.44957794112555,
      "output_token_source": "usage",
      "output_tokens": 205,
      "planned_context_tokens": 4096,
      "prompt_tokens": 2969,
      "reasoning_chunks": 0,
      "request_index": 22,
      "time_to_first_output_ms": 170.25289998855442,
      "ttft_ms": 170.25329999160022,
      "visible_generation_ms": 1927.7134999865666
    },
    {
      "content_chunks": 22,
      "decode_tok_s": 109.16158138957316,
      "e2e_ms": 657.6662000152282,
      "effective_prefill_tok_s": 17246.847755837432,
      "generation_ms": 485.5187999783084,
      "mean_inter_token_latency_ms": 9.160732075062421,
      "output_token_source": "usage",
      "output_tokens": 54,
      "planned_context_tokens": 4096,
      "prompt_tokens": 2969,
      "reasoning_chunks": 0,
      "request_index": 23,
      "time_to_first_output_ms": 172.14740003691986,
      "ttft_ms": 172.14780003996566,
      "visible_generation_ms": 485.5183999752626
    },
    {
      "content_chunks": 48,
      "decode_tok_s": 106.33236498220886,
      "e2e_ms": 1272.1812999807298,
      "effective_prefill_tok_s": 17275.91066714128,
      "generation_ms": 1100.3234999952838,
      "mean_inter_token_latency_ms": 9.40447435893405,
      "output_token_source": "usage",
      "output_tokens": 118,
      "planned_context_tokens": 4096,
      "prompt_tokens": 2969,
      "reasoning_chunks": 0,
      "request_index": 24,
      "time_to_first_output_ms": 171.85779998544604,
      "ttft_ms": 171.85819998849183,
      "visible_generation_ms": 1100.323099992238
    },
    {
      "content_chunks": 17,
      "decode_tok_s": 83.83437895503221,
      "e2e_ms": 542.0405000331812,
      "effective_prefill_tok_s": 17235.193930118723,
      "generation_ms": 369.77670004125684,
      "mean_inter_token_latency_ms": 11.928280646492157,
      "output_token_source": "usage",
      "output_tokens": 32,
      "planned_context_tokens": 4096,
      "prompt_tokens": 2969,
      "reasoning_chunks": 0,
      "request_index": 25,
      "time_to_first_output_ms": 172.26379999192432,
      "ttft_ms": 172.264399996493,
      "visible_generation_ms": 369.77610003668815
    },
    {
      "content_chunks": 93,
      "decode_tok_s": 117.28580771348682,
      "e2e_ms": 2347.314100014046,
      "effective_prefill_tok_s": 17148.170822002096,
      "generation_ms": 2174.1761000012048,
      "mean_inter_token_latency_ms": 8.52618078431845,
      "output_token_source": "usage",
      "output_tokens": 256,
      "planned_context_tokens": 4096,
      "prompt_tokens": 2969,
      "reasoning_chunks": 0,
      "request_index": 26,
      "time_to_first_output_ms": 173.13800001284108,
      "ttft_ms": 173.1385999592021,
      "visible_generation_ms": 2174.1755000548437
    },
    {
      "content_chunks": 15,
      "decode_tok_s": 124.3994314080754,
      "e2e_ms": 505.2140999468975,
      "effective_prefill_tok_s": 16904.799053787345,
      "generation_ms": 329.58349998807535,
      "mean_inter_token_latency_ms": 8.038621950928666,
      "output_token_source": "usage",
      "output_tokens": 42,
      "planned_context_tokens": 4096,
      "prompt_tokens": 2969,
      "reasoning_chunks": 0,
      "request_index": 27,
      "time_to_first_output_ms": 175.63059995882213,
      "ttft_ms": 175.63109996262938,
      "visible_generation_ms": 329.5829999842681
    },
    {
      "content_chunks": 13,
      "decode_tok_s": 133.83273440226702,
      "e2e_ms": 464.67289997963235,
      "effective_prefill_tok_s": 15775.066364216293,
      "generation_ms": 276.4644999988377,
      "mean_inter_token_latency_ms": 7.4720135134821,
      "output_token_source": "usage",
      "output_tokens": 38,
      "planned_context_tokens": 4096,
      "prompt_tokens": 2969,
      "reasoning_chunks": 0,
      "request_index": 28,
      "time_to_first_output_ms": 188.20839998079464,
      "ttft_ms": 188.20899998536333,
      "visible_generation_ms": 276.463899994269
    },
    {
      "content_chunks": 28,
      "decode_tok_s": 104.58005957021076,
      "e2e_ms": 795.5843000090681,
      "effective_prefill_tok_s": 17058.228370866997,
      "generation_ms": 621.5334000298753,
      "mean_inter_token_latency_ms": 9.562052308151928,
      "output_token_source": "usage",
      "output_tokens": 66,
      "planned_context_tokens": 4096,
      "prompt_tokens": 2969,
      "reasoning_chunks": 0,
      "request_index": 29,
      "time_to_first_output_ms": 174.05089997919276,
      "ttft_ms": 174.05129998223856,
      "visible_generation_ms": 621.5330000268295
    },
    {
      "content_chunks": 20,
      "decode_tok_s": 102.79839092336192,
      "e2e_ms": 631.1772000044584,
      "effective_prefill_tok_s": 17066.0038749598,
      "generation_ms": 457.2055999888107,
      "mean_inter_token_latency_ms": 9.727778723166185,
      "output_token_source": "usage",
      "output_tokens": 48,
      "planned_context_tokens": 4096,
      "prompt_tokens": 2969,
      "reasoning_chunks": 0,
      "request_index": 30,
      "time_to_first_output_ms": 173.9716000156477,
      "ttft_ms": 173.9720000186935,
      "visible_generation_ms": 457.2051999857649
    },
    {
      "content_chunks": 23,
      "decode_tok_s": 122.21553103109895,
      "e2e_ms": 690.9954000147991,
      "effective_prefill_tok_s": 16162.669729979985,
      "generation_ms": 507.3005000012927,
      "mean_inter_token_latency_ms": 8.182266129053108,
      "output_token_source": "usage",
      "output_tokens": 63,
      "planned_context_tokens": 4096,
      "prompt_tokens": 2969,
      "reasoning_chunks": 0,
      "request_index": 31,
      "time_to_first_output_ms": 183.6949000135064,
      "ttft_ms": 183.6953000165522,
      "visible_generation_ms": 507.3000999982469
    },
    {
      "content_chunks": 23,
      "decode_tok_s": 104.67109576307456,
      "e2e_ms": 677.2370000253431,
      "effective_prefill_tok_s": 17373.85086216434,
      "generation_ms": 506.34800002444535,
      "mean_inter_token_latency_ms": 9.553735849517837,
      "output_token_source": "usage",
      "output_tokens": 54,
      "planned_context_tokens": 4096,
      "prompt_tokens": 2969,
      "reasoning_chunks": 0,
      "request_index": 32,
      "time_to_first_output_ms": 170.8890000008978,
      "ttft_ms": 170.88950000470504,
      "visible_generation_ms": 506.3475000206381
    },
    {
      "content_chunks": 3,
      "decode_tok_s": 59.172210359557,
      "e2e_ms": 240.68809999153018,
      "effective_prefill_tok_s": 17153.045143389998,
      "generation_ms": 67.59930000407621,
      "mean_inter_token_latency_ms": 16.899825001019053,
      "output_token_source": "usage",
      "output_tokens": 5,
      "planned_context_tokens": 4096,
      "prompt_tokens": 2969,
      "reasoning_chunks": 0,
      "request_index": 33,
      "time_to_first_output_ms": 173.08879998745397,
      "ttft_ms": 173.0892999912612,
      "visible_generation_ms": 67.59880000026897
    },
    {
      "content_chunks": 17,
      "decode_tok_s": 102.47515248191662,
      "e2e_ms": 545.7918999600224,
      "effective_prefill_tok_s": 16968.594099022444,
      "generation_ms": 370.82159996498376,
      "mean_inter_token_latency_ms": 9.758463156973257,
      "output_token_source": "usage",
      "output_tokens": 39,
      "planned_context_tokens": 4096,
      "prompt_tokens": 2969,
      "reasoning_chunks": 0,
      "request_index": 34,
      "time_to_first_output_ms": 174.97029999503866,
      "ttft_ms": 174.97069999808446,
      "visible_generation_ms": 370.82119996193796
    },
    {
      "content_chunks": 9,
      "decode_tok_s": 102.05523123286312,
      "e2e_ms": 355.1844999892637,
      "effective_prefill_tok_s": 17566.92471532026,
      "generation_ms": 186.17369997082278,
      "mean_inter_token_latency_ms": 9.798615787938042,
      "output_token_source": "usage",
      "output_tokens": 20,
      "planned_context_tokens": 4096,
      "prompt_tokens": 2969,
      "reasoning_chunks": 0,
      "request_index": 35,
      "time_to_first_output_ms": 169.01080001844093,
      "ttft_ms": 169.01109996251762,
      "visible_generation_ms": 186.1734000267461
    },
    {
      "content_chunks": 32,
      "decode_tok_s": 104.33320176643896,
      "e2e_ms": 907.0057999924757,
      "effective_prefill_tok_s": 17569.53398995248,
      "generation_ms": 738.0200999905355,
      "mean_inter_token_latency_ms": 9.584676623253708,
      "output_token_source": "usage",
      "output_tokens": 78,
      "planned_context_tokens": 4096,
      "prompt_tokens": 2969,
      "reasoning_chunks": 0,
      "request_index": 36,
      "time_to_first_output_ms": 168.98570000194013,
      "ttft_ms": 168.98600000422448,
      "visible_generation_ms": 738.0197999882512
    },
    {
      "content_chunks": 4,
      "decode_tok_s": 44.190105829543796,
      "e2e_ms": 264.19620000524446,
      "effective_prefill_tok_s": 17094.834009749247,
      "generation_ms": 90.518000011798,
      "mean_inter_token_latency_ms": 22.6295000029495,
      "output_token_source": "usage",
      "output_tokens": 5,
      "planned_context_tokens": 4096,
      "prompt_tokens": 2969,
      "reasoning_chunks": 0,
      "request_index": 37,
      "time_to_first_output_ms": 173.67819999344647,
      "ttft_ms": 173.67879999801517,
      "visible_generation_ms": 90.5174000072293
    },
    {
      "content_chunks": 23,
      "decode_tok_s": 99.86631620236801,
      "e2e_ms": 682.8738999902271,
      "effective_prefill_tok_s": 17242.460710177395,
      "generation_ms": 510.682700027246,
      "mean_inter_token_latency_ms": 10.01338627504404,
      "output_token_source": "usage",
      "output_tokens": 52,
      "planned_context_tokens": 4096,
      "prompt_tokens": 2969,
      "reasoning_chunks": 0,
      "request_index": 38,
      "time_to_first_output_ms": 172.19119996298105,
      "ttft_ms": 172.19159996602684,
      "visible_generation_ms": 510.68230002420023
    },
    {
      "content_chunks": 3,
      "decode_tok_s": 57.271554529316084,
      "e2e_ms": 245.16059999587014,
      "effective_prefill_tok_s": 16934.95073588665,
      "generation_ms": 69.84269997337833,
      "mean_inter_token_latency_ms": 17.460674993344583,
      "output_token_source": "usage",
      "output_tokens": 5,
      "planned_context_tokens": 4096,
      "prompt_tokens": 2969,
      "reasoning_chunks": 0,
      "request_index": 39,
      "time_to_first_output_ms": 175.3179000224918,
      "ttft_ms": 175.3183999680914,
      "visible_generation_ms": 69.84220002777874
    },
    {
      "content_chunks": 16,
      "decode_tok_s": 107.01170899517564,
      "e2e_ms": 540.9160000272095,
      "effective_prefill_tok_s": 16824.398948009548,
      "generation_ms": 364.44610002217814,
      "mean_inter_token_latency_ms": 9.344771795440465,
      "output_token_source": "usage",
      "output_tokens": 40,
      "planned_context_tokens": 4096,
      "prompt_tokens": 2969,
      "reasoning_chunks": 0,
      "request_index": 40,
      "time_to_first_output_ms": 176.46990000503138,
      "ttft_ms": 176.47030000807717,
      "visible_generation_ms": 364.44570001913235
    },
    {
      "content_chunks": 16,
      "decode_tok_s": 113.90902174857922,
      "e2e_ms": 519.9312000186183,
      "effective_prefill_tok_s": 16721.79583462019,
      "generation_ms": 342.37849997589365,
      "mean_inter_token_latency_ms": 8.778935896817785,
      "output_token_source": "usage",
      "output_tokens": 40,
      "planned_context_tokens": 4096,
      "prompt_tokens": 2969,
      "reasoning_chunks": 0,
      "request_index": 41,
      "time_to_first_output_ms": 177.5527000427246,
      "ttft_ms": 177.5531000457704,
      "visible_generation_ms": 342.37809997284785
    },
    {
      "content_chunks": 17,
      "decode_tok_s": 109.70829075538767,
      "e2e_ms": 565.4250999796204,
      "effective_prefill_tok_s": 17114.700198662264,
      "generation_ms": 391.94850000785664,
      "mean_inter_token_latency_ms": 9.115081395531549,
      "output_token_source": "usage",
      "output_tokens": 44,
      "planned_context_tokens": 4096,
      "prompt_tokens": 2969,
      "reasoning_chunks": 0,
      "request_index": 42,
      "time_to_first_output_ms": 173.4765999717638,
      "ttft_ms": 173.47689997404814,
      "visible_generation_ms": 391.9482000055723
    },
    {
      "content_chunks": 21,
      "decode_tok_s": 109.8718396810057,
      "e2e_ms": 656.6477000014856,
      "effective_prefill_tok_s": 17037.026410676295,
      "generation_ms": 482.3802000028081,
      "mean_inter_token_latency_ms": 9.101513207600153,
      "output_token_source": "usage",
      "output_tokens": 54,
      "planned_context_tokens": 4096,
      "prompt_tokens": 2969,
      "reasoning_chunks": 0,
      "request_index": 43,
      "time_to_first_output_ms": 174.26749999867752,
      "ttft_ms": 174.26800000248477,
      "visible_generation_ms": 482.3796999990009
    },
    {
      "content_chunks": 22,
      "decode_tok_s": 91.69911685120412,
      "e2e_ms": 674.8640000005253,
      "effective_prefill_tok_s": 17139.716693046583,
      "generation_ms": 501.6406000358984,
      "mean_inter_token_latency_ms": 10.90523043556301,
      "output_token_source": "usage",
      "output_tokens": 47,
      "planned_context_tokens": 4096,
      "prompt_tokens": 2969,
      "reasoning_chunks": 0,
      "request_index": 44,
      "time_to_first_output_ms": 173.22339996462688,
      "ttft_ms": 173.22409996995702,
      "visible_generation_ms": 501.63990003056824
    },
    {
      "content_chunks": 21,
      "decode_tok_s": 110.48668760646771,
      "e2e_ms": 654.6293999999762,
      "effective_prefill_tok_s": 16972.15400522003,
      "generation_ms": 479.6957999933511,
      "mean_inter_token_latency_ms": 9.050864150817945,
      "output_token_source": "usage",
      "output_tokens": 54,
      "planned_context_tokens": 4096,
      "prompt_tokens": 2969,
      "reasoning_chunks": 0,
      "request_index": 45,
      "time_to_first_output_ms": 174.93360000662506,
      "ttft_ms": 174.9339000089094,
      "visible_generation_ms": 479.69549999106675
    },
    {
      "content_chunks": 12,
      "decode_tok_s": 121.19548793704676,
      "e2e_ms": 427.82380001153797,
      "effective_prefill_tok_s": 17257.744911950285,
      "generation_ms": 255.78509998740628,
      "mean_inter_token_latency_ms": 8.251132257658266,
      "output_token_source": "usage",
      "output_tokens": 32,
      "planned_context_tokens": 4096,
      "prompt_tokens": 2969,
      "reasoning_chunks": 0,
      "request_index": 46,
      "time_to_first_output_ms": 172.03870002413169,
      "ttft_ms": 172.03920002793893,
      "visible_generation_ms": 255.78459998359904
    },
    {
      "content_chunks": 14,
      "decode_tok_s": 100.63169870795565,
      "e2e_ms": 468.97980000358075,
      "effective_prefill_tok_s": 17376.494620779264,
      "generation_ms": 298.1168000260368,
      "mean_inter_token_latency_ms": 9.93722666753456,
      "output_token_source": "usage",
      "output_tokens": 31,
      "planned_context_tokens": 4096,
      "prompt_tokens": 2969,
      "reasoning_chunks": 0,
      "request_index": 47,
      "time_to_first_output_ms": 170.86299997754395,
      "ttft_ms": 170.86359998211265,
      "visible_generation_ms": 298.1162000214681
    },
    {
      "content_chunks": 33,
      "decode_tok_s": 105.16306968014047,
      "e2e_ms": 908.5706000332721,
      "effective_prefill_tok_s": 16833.508714521242,
      "generation_ms": 732.1961999987252,
      "mean_inter_token_latency_ms": 9.509041558425002,
      "output_token_source": "usage",
      "output_tokens": 78,
      "planned_context_tokens": 4096,
      "prompt_tokens": 2969,
      "reasoning_chunks": 0,
      "request_index": 48,
      "time_to_first_output_ms": 176.37440003454685,
      "ttft_ms": 176.374799979385,
      "visible_generation_ms": 732.1958000538871
    },
    {
      "content_chunks": 19,
      "decode_tok_s": 95.05120646451586,
      "e2e_ms": 596.5528999804519,
      "effective_prefill_tok_s": 16895.51583154961,
      "generation_ms": 420.8257999853231,
      "mean_inter_token_latency_ms": 10.520644999633078,
      "output_token_source": "usage",
      "output_tokens": 41,
      "planned_context_tokens": 4096,
      "prompt_tokens": 2969,
      "reasoning_chunks": 0,
      "request_index": 49,
      "time_to_first_output_ms": 175.7270999951288,
      "ttft_ms": 175.72739999741316,
      "visible_generation_ms": 420.82549998303875
    },
    {
      "content_chunks": 15,
      "decode_tok_s": 115.1273942206176,
      "e2e_ms": 517.778699984774,
      "effective_prefill_tok_s": 17430.102004147575,
      "generation_ms": 347.4411999923177,
      "mean_inter_token_latency_ms": 8.686029999807943,
      "output_token_source": "usage",
      "output_tokens": 41,
      "planned_context_tokens": 4096,
      "prompt_tokens": 2969,
      "reasoning_chunks": 0,
      "request_index": 50,
      "time_to_first_output_ms": 170.3374999924563,
      "ttft_ms": 170.33799999626353,
      "visible_generation_ms": 347.44069998851046
    },
    {
      "content_chunks": 15,
      "decode_tok_s": 101.45839190552084,
      "e2e_ms": 522.3577999859117,
      "effective_prefill_tok_s": 16737.246098336145,
      "generation_ms": 344.96900002704933,
      "mean_inter_token_latency_ms": 9.85625714362998,
      "output_token_source": "usage",
      "output_tokens": 36,
      "planned_context_tokens": 4096,
      "prompt_tokens": 2969,
      "reasoning_chunks": 0,
      "request_index": 51,
      "time_to_first_output_ms": 177.38879995886236,
      "ttft_ms": 177.38919996190816,
      "visible_generation_ms": 344.96860002400354
    },
    {
      "content_chunks": 24,
      "decode_tok_s": 106.11444708635413,
      "e2e_ms": 698.2392999925651,
      "effective_prefill_tok_s": 17412.754416993404,
      "generation_ms": 527.7320999884978,
      "mean_inter_token_latency_ms": 9.423787499794603,
      "output_token_source": "usage",
      "output_tokens": 57,
      "planned_context_tokens": 4096,
      "prompt_tokens": 2969,
      "reasoning_chunks": 0,
      "request_index": 52,
      "time_to_first_output_ms": 170.50720000406727,
      "ttft_ms": 170.50770000787452,
      "visible_generation_ms": 527.7315999846905
    },
    {
      "content_chunks": 23,
      "decode_tok_s": 115.80177924785147,
      "e2e_ms": 698.5168000101112,
      "effective_prefill_tok_s": 17286.291020727826,
      "generation_ms": 526.7621999955736,
      "mean_inter_token_latency_ms": 8.635445901566781,
      "output_token_source": "usage",
      "output_tokens": 62,
      "planned_context_tokens": 4096,
      "prompt_tokens": 2969,
      "reasoning_chunks": 0,
      "request_index": 53,
      "time_to_first_output_ms": 171.75460001453757,
      "ttft_ms": 171.75500001758337,
      "visible_generation_ms": 526.7617999925278
    },
    {
      "content_chunks": 23,
      "decode_tok_s": 81.818221602559,
      "e2e_ms": 700.4916000296362,
      "effective_prefill_tok_s": 16971.89205178475,
      "generation_ms": 525.555300002452,
      "mean_inter_token_latency_ms": 12.22221627912679,
      "output_token_source": "usage",
      "output_tokens": 44,
      "planned_context_tokens": 4096,
      "prompt_tokens": 2969,
      "reasoning_chunks": 0,
      "request_index": 54,
      "time_to_first_output_ms": 174.9363000271842,
      "ttft_ms": 174.93670003022999,
      "visible_generation_ms": 525.5548999994062
    },
    {
      "content_chunks": 19,
      "decode_tok_s": 92.44216009004008,
      "e2e_ms": 610.0675000343472,
      "effective_prefill_tok_s": 16739.539196899612,
      "generation_ms": 432.70300002768636,
      "mean_inter_token_latency_ms": 10.817575000692159,
      "output_token_source": "usage",
      "output_tokens": 41,
      "planned_context_tokens": 4096,
      "prompt_tokens": 2969,
      "reasoning_chunks": 0,
      "request_index": 55,
      "time_to_first_output_ms": 177.36450000666082,
      "ttft_ms": 177.36480000894517,
      "visible_generation_ms": 432.702700025402
    },
    {
      "content_chunks": 59,
      "decode_tok_s": 114.35324277181738,
      "e2e_ms": 1528.1521000433713,
      "effective_prefill_tok_s": 17191.363205527065,
      "generation_ms": 1355.4491000249982,
      "mean_inter_token_latency_ms": 8.744832903387085,
      "output_token_source": "usage",
      "output_tokens": 156,
      "planned_context_tokens": 4096,
      "prompt_tokens": 2969,
      "reasoning_chunks": 0,
      "request_index": 56,
      "time_to_first_output_ms": 172.70300001837313,
      "ttft_ms": 172.70330002065748,
      "visible_generation_ms": 1355.4488000227138
    },
    {
      "content_chunks": 20,
      "decode_tok_s": 106.21332108240746,
      "e2e_ms": 619.2234000191092,
      "effective_prefill_tok_s": 16800.80716581521,
      "generation_ms": 442.5057000480592,
      "mean_inter_token_latency_ms": 9.415014894639558,
      "output_token_source": "usage",
      "output_tokens": 48,
      "planned_context_tokens": 4096,
      "prompt_tokens": 2969,
      "reasoning_chunks": 0,
      "request_index": 57,
      "time_to_first_output_ms": 176.71769997105002,
      "ttft_ms": 176.71809997409582,
      "visible_generation_ms": 442.5053000450134
    },
    {
      "content_chunks": 17,
      "decode_tok_s": 91.39404138331925,
      "e2e_ms": 554.6049000113271,
      "effective_prefill_tok_s": 16260.527716384178,
      "generation_ms": 372.0154999755323,
      "mean_inter_token_latency_ms": 10.941632352221538,
      "output_token_source": "usage",
      "output_tokens": 35,
      "planned_context_tokens": 4096,
      "prompt_tokens": 2969,
      "reasoning_chunks": 0,
      "request_index": 58,
      "time_to_first_output_ms": 182.58940003579482,
      "ttft_ms": 182.58970003807917,
      "visible_generation_ms": 372.01519997324795
    },
    {
      "content_chunks": 18,
      "decode_tok_s": 111.79503700812305,
      "e2e_ms": 570.9616999956779,
      "effective_prefill_tok_s": 16737.670697394344,
      "generation_ms": 393.5774000128731,
      "mean_inter_token_latency_ms": 8.94494090938348,
      "output_token_source": "usage",
      "output_tokens": 45,
      "planned_context_tokens": 4096,
      "prompt_tokens": 2969,
      "reasoning_chunks": 0,
      "request_index": 59,
      "time_to_first_output_ms": 177.3842999828048,
      "ttft_ms": 177.3846999858506,
      "visible_generation_ms": 393.5770000098273
    }
  ],
  "requests": 60,
  "run_id": "benchmark-20260902T113958Z",
  "schema": "anvil-serving.benchmark/v1",
  "serve_flags": {
    "no_thinking": false,
    "reasoning_effort": null,
    "shared_prefix_burst": false,
    "thinking_mode": "disabled"
  },
  "source_recipe": null,
  "started_at": "2026-09-02T11:39:58Z",
  "tier": null,
  "timing_methodology": {
    "clock": "client time.perf_counter",
    "decode": "usage completion tokens after the first token divided by client-observed generation time; completion tokens may include reasoning tokens",
    "effective_prefill": "usage.prompt_tokens divided by client-observed time to first output; includes queueing, scheduling, prefill, and first-token work",
    "generation": "client-observed E2E minus time to first output",
    "mean_inter_token_latency": "client-observed generation time divided by completion tokens after the first token; not a raw per-token timestamp trace",
    "time_to_first_output": "request start through the first non-empty streamed reasoning or content delta",
    "ttft": "request start through the first non-empty streamed content delta"
  },
  "wall_clock_ms": 43739.36760000652
}
