{
  "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": "2026-09-02T11:21:43Z",
  "gpu": "2x RTX PRO 6000 Blackwell Max-Q TP2 WSL2",
  "manifest": null,
  "max_context_tokens": 393216,
  "max_tokens": 256,
  "measurement_protocol": "capacity-v3",
  "metrics": {
    "content_chunks": 1964,
    "content_chunks_s": null,
    "decode_tok_s_p50": 102.19275206464441,
    "decode_tok_s_p95": 119.09889474589053,
    "e2e_p50_ms": 645.7104999572039,
    "e2e_p95_ms": 2363.374100008514,
    "effective_prefill_tok_s_p50": 16141.650734953235,
    "effective_prefill_tok_s_p95": 17324.742343993166,
    "generation_p50_ms": 458.085399994161,
    "generation_p95_ms": 2172.676699992735,
    "mean_inter_token_latency_ms_p50": 9.777266799937934,
    "mean_inter_token_latency_ms_p95": 16.657999993185513,
    "output_token_sources": {
      "usage": 60
    },
    "output_tokens": 4808,
    "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": 85.44527147280385,
    "time_to_first_output_p50_ms": 183.42039996059611,
    "time_to_first_output_p95_ms": 192.8695000242442,
    "ttft_p50_ms": 183.42089996440336,
    "ttft_p95_ms": 192.86980002652854,
    "ttft_samples": 60
  },
  "model": "glm53-flash-ormandj-sglang-sm120-tp2-393k-c1-adaptive-mtp",
  "request_timings": [
    {
      "content_chunks": 23,
      "decode_tok_s": 95.60187132366254,
      "e2e_ms": 1139.5114999613725,
      "effective_prefill_tok_s": 5166.45234044311,
      "generation_ms": 564.8424999671988,
      "mean_inter_token_latency_ms": 10.460046295688867,
      "output_token_source": "usage",
      "output_tokens": 55,
      "planned_context_tokens": 4096,
      "prompt_tokens": 2969,
      "reasoning_chunks": 0,
      "request_index": 0,
      "time_to_first_output_ms": 574.6689999941736,
      "ttft_ms": 574.6695999987423,
      "visible_generation_ms": 564.8418999626301
    },
    {
      "content_chunks": 17,
      "decode_tok_s": 105.03904624977368,
      "e2e_ms": 553.3399999840185,
      "effective_prefill_tok_s": 16308.751191362617,
      "generation_ms": 371.2904999847524,
      "mean_inter_token_latency_ms": 9.520269230378267,
      "output_token_source": "usage",
      "output_tokens": 40,
      "planned_context_tokens": 4096,
      "prompt_tokens": 2969,
      "reasoning_chunks": 0,
      "request_index": 1,
      "time_to_first_output_ms": 182.04949999926612,
      "ttft_ms": 182.05009994562715,
      "visible_generation_ms": 371.28990003839135
    },
    {
      "content_chunks": 28,
      "decode_tok_s": 112.69706289662942,
      "e2e_ms": 810.122599999886,
      "effective_prefill_tok_s": 15709.958914694853,
      "generation_ms": 621.1341999587603,
      "mean_inter_token_latency_ms": 8.873345713696576,
      "output_token_source": "usage",
      "output_tokens": 71,
      "planned_context_tokens": 4096,
      "prompt_tokens": 2969,
      "reasoning_chunks": 0,
      "request_index": 2,
      "time_to_first_output_ms": 188.98840004112571,
      "ttft_ms": 188.9886999852024,
      "visible_generation_ms": 621.1339000146836
    },
    {
      "content_chunks": 29,
      "decode_tok_s": 108.48311623456472,
      "e2e_ms": 843.1493999669328,
      "effective_prefill_tok_s": 15003.459033113735,
      "generation_ms": 645.2616999740712,
      "mean_inter_token_latency_ms": 9.218024285343875,
      "output_token_source": "usage",
      "output_tokens": 71,
      "planned_context_tokens": 4096,
      "prompt_tokens": 2969,
      "reasoning_chunks": 0,
      "request_index": 3,
      "time_to_first_output_ms": 197.8876999928616,
      "ttft_ms": 197.88799999514595,
      "visible_generation_ms": 645.2613999717869
    },
    {
      "content_chunks": 106,
      "decode_tok_s": 102.27807223245131,
      "e2e_ms": 2623.07700002566,
      "effective_prefill_tok_s": 16608.833165507713,
      "generation_ms": 2444.3166999844834,
      "mean_inter_token_latency_ms": 9.777266799937934,
      "output_token_source": "usage",
      "output_tokens": 251,
      "planned_context_tokens": 4096,
      "prompt_tokens": 2969,
      "reasoning_chunks": 0,
      "request_index": 4,
      "time_to_first_output_ms": 178.76030004117638,
      "ttft_ms": 178.76080004498363,
      "visible_generation_ms": 2444.316199980676
    },
    {
      "content_chunks": 20,
      "decode_tok_s": 102.19275206464441,
      "e2e_ms": 645.7104999572039,
      "effective_prefill_tok_s": 15979.952130123276,
      "generation_ms": 459.91519995732233,
      "mean_inter_token_latency_ms": 9.785429786326008,
      "output_token_source": "usage",
      "output_tokens": 48,
      "planned_context_tokens": 4096,
      "prompt_tokens": 2969,
      "reasoning_chunks": 0,
      "request_index": 5,
      "time_to_first_output_ms": 185.79529999988154,
      "ttft_ms": 185.79560000216588,
      "visible_generation_ms": 459.914899955038
    },
    {
      "content_chunks": 21,
      "decode_tok_s": 99.23063184018369,
      "e2e_ms": 676.5910999965854,
      "effective_prefill_tok_s": 15393.828467574132,
      "generation_ms": 483.7215999723412,
      "mean_inter_token_latency_ms": 10.077533332757108,
      "output_token_source": "usage",
      "output_tokens": 49,
      "planned_context_tokens": 4096,
      "prompt_tokens": 2969,
      "reasoning_chunks": 0,
      "request_index": 6,
      "time_to_first_output_ms": 192.8695000242442,
      "ttft_ms": 192.86980002652854,
      "visible_generation_ms": 483.72129997005686
    },
    {
      "content_chunks": 95,
      "decode_tok_s": 105.39994284504719,
      "e2e_ms": 2363.374100008514,
      "effective_prefill_tok_s": 15569.168744588738,
      "generation_ms": 2172.676699992735,
      "mean_inter_token_latency_ms": 9.487671179007576,
      "output_token_source": "usage",
      "output_tokens": 230,
      "planned_context_tokens": 4096,
      "prompt_tokens": 2969,
      "reasoning_chunks": 0,
      "request_index": 7,
      "time_to_first_output_ms": 190.69740001577884,
      "ttft_ms": 190.69760001730174,
      "visible_generation_ms": 2172.676499991212
    },
    {
      "content_chunks": 17,
      "decode_tok_s": 95.38294753245927,
      "e2e_ms": 538.9742000261322,
      "effective_prefill_tok_s": 17258.38694492008,
      "generation_ms": 366.94190005073324,
      "mean_inter_token_latency_ms": 10.484054287163806,
      "output_token_source": "usage",
      "output_tokens": 36,
      "planned_context_tokens": 4096,
      "prompt_tokens": 2969,
      "reasoning_chunks": 0,
      "request_index": 8,
      "time_to_first_output_ms": 172.03229997539893,
      "ttft_ms": 172.03259997768328,
      "visible_generation_ms": 366.9416000484489
    },
    {
      "content_chunks": 19,
      "decode_tok_s": 107.93028071346608,
      "e2e_ms": 606.5032000187784,
      "effective_prefill_tok_s": 17358.827245074644,
      "generation_ms": 435.46629999764264,
      "mean_inter_token_latency_ms": 9.265240425481759,
      "output_token_source": "usage",
      "output_tokens": 48,
      "planned_context_tokens": 4096,
      "prompt_tokens": 2969,
      "reasoning_chunks": 0,
      "request_index": 9,
      "time_to_first_output_ms": 171.0369000211358,
      "ttft_ms": 171.0373000241816,
      "visible_generation_ms": 435.46589999459684
    },
    {
      "content_chunks": 83,
      "decode_tok_s": 102.97200439486828,
      "e2e_ms": 2072.5352999870665,
      "effective_prefill_tok_s": 16603.594630241125,
      "generation_ms": 1893.718599982094,
      "mean_inter_token_latency_ms": 9.71137743580561,
      "output_token_source": "usage",
      "output_tokens": 196,
      "planned_context_tokens": 4096,
      "prompt_tokens": 2969,
      "reasoning_chunks": 0,
      "request_index": 10,
      "time_to_first_output_ms": 178.81670000497252,
      "ttft_ms": 178.81720000877976,
      "visible_generation_ms": 1893.7180999782868
    },
    {
      "content_chunks": 70,
      "decode_tok_s": 98.15782610372136,
      "e2e_ms": 1793.981500028167,
      "effective_prefill_tok_s": 16107.078162139494,
      "generation_ms": 1609.6526000183076,
      "mean_inter_token_latency_ms": 10.187674683660175,
      "output_token_source": "usage",
      "output_tokens": 159,
      "planned_context_tokens": 4096,
      "prompt_tokens": 2969,
      "reasoning_chunks": 0,
      "request_index": 11,
      "time_to_first_output_ms": 184.32890000985935,
      "ttft_ms": 184.32950001442805,
      "visible_generation_ms": 1609.6520000137389
    },
    {
      "content_chunks": 16,
      "decode_tok_s": 100.34594944001934,
      "e2e_ms": 554.3918000184931,
      "effective_prefill_tok_s": 15990.960177994904,
      "generation_ms": 368.724400002975,
      "mean_inter_token_latency_ms": 9.965524324404731,
      "output_token_source": "usage",
      "output_tokens": 38,
      "planned_context_tokens": 4096,
      "prompt_tokens": 2969,
      "reasoning_chunks": 0,
      "request_index": 12,
      "time_to_first_output_ms": 185.66740001551807,
      "ttft_ms": 185.66770001780242,
      "visible_generation_ms": 368.72410000069067
    },
    {
      "content_chunks": 25,
      "decode_tok_s": 115.45084003223465,
      "e2e_ms": 754.4698000419885,
      "effective_prefill_tok_s": 14836.001964514991,
      "generation_ms": 554.3485000380315,
      "mean_inter_token_latency_ms": 8.661695313094242,
      "output_token_source": "usage",
      "output_tokens": 65,
      "planned_context_tokens": 4096,
      "prompt_tokens": 2969,
      "reasoning_chunks": 0,
      "request_index": 13,
      "time_to_first_output_ms": 200.121300003957,
      "ttft_ms": 200.1217000070028,
      "visible_generation_ms": 554.3481000349857
    },
    {
      "content_chunks": 18,
      "decode_tok_s": 106.84609039252345,
      "e2e_ms": 588.1099000107497,
      "effective_prefill_tok_s": 16840.364236279267,
      "generation_ms": 411.8072999990545,
      "mean_inter_token_latency_ms": 9.35925681816033,
      "output_token_source": "usage",
      "output_tokens": 45,
      "planned_context_tokens": 4096,
      "prompt_tokens": 2969,
      "reasoning_chunks": 0,
      "request_index": 14,
      "time_to_first_output_ms": 176.30260001169518,
      "ttft_ms": 176.30280001321808,
      "visible_generation_ms": 411.8070999975316
    },
    {
      "content_chunks": 43,
      "decode_tok_s": 101.04510956844219,
      "e2e_ms": 1163.8372999732383,
      "effective_prefill_tok_s": 17045.555670914073,
      "generation_ms": 989.6569999982603,
      "mean_inter_token_latency_ms": 9.896569999982603,
      "output_token_source": "usage",
      "output_tokens": 101,
      "planned_context_tokens": 4096,
      "prompt_tokens": 2969,
      "reasoning_chunks": 0,
      "request_index": 15,
      "time_to_first_output_ms": 174.18029997497797,
      "ttft_ms": 174.18069997802377,
      "visible_generation_ms": 989.6565999952145
    },
    {
      "content_chunks": 81,
      "decode_tok_s": 107.89438708845644,
      "e2e_ms": 2006.335999991279,
      "effective_prefill_tok_s": 17341.469242655337,
      "generation_ms": 1835.1279000053182,
      "mean_inter_token_latency_ms": 9.268322727299587,
      "output_token_source": "usage",
      "output_tokens": 199,
      "planned_context_tokens": 4096,
      "prompt_tokens": 2969,
      "reasoning_chunks": 0,
      "request_index": 16,
      "time_to_first_output_ms": 171.20809998596087,
      "ttft_ms": 171.20849998900667,
      "visible_generation_ms": 1835.1275000022724
    },
    {
      "content_chunks": 40,
      "decode_tok_s": 98.69555874745814,
      "e2e_ms": 1074.6076999930665,
      "effective_prefill_tok_s": 16226.091925870676,
      "generation_ms": 891.6307999752462,
      "mean_inter_token_latency_ms": 10.132168181536889,
      "output_token_source": "usage",
      "output_tokens": 89,
      "planned_context_tokens": 4096,
      "prompt_tokens": 2969,
      "reasoning_chunks": 0,
      "request_index": 17,
      "time_to_first_output_ms": 182.9769000178203,
      "ttft_ms": 182.97720002010465,
      "visible_generation_ms": 891.6304999729618
    },
    {
      "content_chunks": 16,
      "decode_tok_s": 119.37816207994453,
      "e2e_ms": 531.6380999865942,
      "effective_prefill_tok_s": 15776.466228607462,
      "generation_ms": 343.4463999583386,
      "mean_inter_token_latency_ms": 8.376741462398503,
      "output_token_source": "usage",
      "output_tokens": 42,
      "planned_context_tokens": 4096,
      "prompt_tokens": 2969,
      "reasoning_chunks": 0,
      "request_index": 18,
      "time_to_first_output_ms": 188.19170002825558,
      "ttft_ms": 188.19210003130138,
      "visible_generation_ms": 343.4459999552928
    },
    {
      "content_chunks": 14,
      "decode_tok_s": 113.67534455938583,
      "e2e_ms": 488.5869999998249,
      "effective_prefill_tok_s": 15668.406074805023,
      "generation_ms": 299.09739998402074,
      "mean_inter_token_latency_ms": 8.796982352471199,
      "output_token_source": "usage",
      "output_tokens": 35,
      "planned_context_tokens": 4096,
      "prompt_tokens": 2969,
      "reasoning_chunks": 0,
      "request_index": 19,
      "time_to_first_output_ms": 189.48960001580417,
      "ttft_ms": 189.49029996292666,
      "visible_generation_ms": 299.09670003689826
    },
    {
      "content_chunks": 15,
      "decode_tok_s": 105.43237194115555,
      "e2e_ms": 497.5588000379503,
      "effective_prefill_tok_s": 16958.23328280154,
      "generation_ms": 322.48160004382953,
      "mean_inter_token_latency_ms": 9.484752942465574,
      "output_token_source": "usage",
      "output_tokens": 35,
      "planned_context_tokens": 4096,
      "prompt_tokens": 2969,
      "reasoning_chunks": 0,
      "request_index": 20,
      "time_to_first_output_ms": 175.07719999412075,
      "ttft_ms": 175.07759999716654,
      "visible_generation_ms": 322.48120004078373
    },
    {
      "content_chunks": 14,
      "decode_tok_s": 112.53921589476116,
      "e2e_ms": 484.68519997550175,
      "effective_prefill_tok_s": 17094.410781439372,
      "generation_ms": 311.0027000075206,
      "mean_inter_token_latency_ms": 8.885791428786304,
      "output_token_source": "usage",
      "output_tokens": 36,
      "planned_context_tokens": 4096,
      "prompt_tokens": 2969,
      "reasoning_chunks": 0,
      "request_index": 21,
      "time_to_first_output_ms": 173.68249996798113,
      "ttft_ms": 173.68289997102693,
      "visible_generation_ms": 311.0023000044748
    },
    {
      "content_chunks": 20,
      "decode_tok_s": 97.24003386466237,
      "e2e_ms": 619.6949000004679,
      "effective_prefill_tok_s": 16727.684116963548,
      "generation_ms": 442.20469996798784,
      "mean_inter_token_latency_ms": 10.28383023181367,
      "output_token_source": "usage",
      "output_tokens": 44,
      "planned_context_tokens": 4096,
      "prompt_tokens": 2969,
      "reasoning_chunks": 0,
      "request_index": 22,
      "time_to_first_output_ms": 177.49020003248006,
      "ttft_ms": 177.4905000347644,
      "visible_generation_ms": 442.2043999657035
    },
    {
      "content_chunks": 3,
      "decode_tok_s": 57.8340567566008,
      "e2e_ms": 240.53680000361055,
      "effective_prefill_tok_s": 17324.742343993166,
      "generation_ms": 69.1633999813348,
      "mean_inter_token_latency_ms": 17.2908499953337,
      "output_token_source": "usage",
      "output_tokens": 5,
      "planned_context_tokens": 4096,
      "prompt_tokens": 2969,
      "reasoning_chunks": 0,
      "request_index": 23,
      "time_to_first_output_ms": 171.37340002227575,
      "ttft_ms": 171.37380002532154,
      "visible_generation_ms": 69.16299997828901
    },
    {
      "content_chunks": 20,
      "decode_tok_s": 84.58991267021479,
      "e2e_ms": 622.6384000037797,
      "effective_prefill_tok_s": 16028.374922532032,
      "generation_ms": 437.404400028754,
      "mean_inter_token_latency_ms": 11.821740541317677,
      "output_token_source": "usage",
      "output_tokens": 38,
      "planned_context_tokens": 4096,
      "prompt_tokens": 2969,
      "reasoning_chunks": 0,
      "request_index": 24,
      "time_to_first_output_ms": 185.23399997502565,
      "ttft_ms": 185.23459997959435,
      "visible_generation_ms": 437.40380002418533
    },
    {
      "content_chunks": 18,
      "decode_tok_s": 96.2646498396104,
      "e2e_ms": 611.5445999894291,
      "effective_prefill_tok_s": 15993.716715265895,
      "generation_ms": 425.9091999847442,
      "mean_inter_token_latency_ms": 10.38802926792059,
      "output_token_source": "usage",
      "output_tokens": 42,
      "planned_context_tokens": 4096,
      "prompt_tokens": 2969,
      "reasoning_chunks": 0,
      "request_index": 25,
      "time_to_first_output_ms": 185.63540000468493,
      "ttft_ms": 185.63570000696927,
      "visible_generation_ms": 425.90889998245984
    },
    {
      "content_chunks": 71,
      "decode_tok_s": 118.43394089088979,
      "e2e_ms": 1872.6392000098713,
      "effective_prefill_tok_s": 16141.650734953235,
      "generation_ms": 1688.7051000376232,
      "mean_inter_token_latency_ms": 8.443525500188116,
      "output_token_source": "usage",
      "output_tokens": 201,
      "planned_context_tokens": 4096,
      "prompt_tokens": 2969,
      "reasoning_chunks": 0,
      "request_index": 26,
      "time_to_first_output_ms": 183.93409997224808,
      "ttft_ms": 183.93449997529387,
      "visible_generation_ms": 1688.7047000345774
    },
    {
      "content_chunks": 29,
      "decode_tok_s": 119.09889474589053,
      "e2e_ms": 835.5585000244901,
      "effective_prefill_tok_s": 17237.155153565353,
      "generation_ms": 663.3143000071868,
      "mean_inter_token_latency_ms": 8.39638354439477,
      "output_token_source": "usage",
      "output_tokens": 80,
      "planned_context_tokens": 4096,
      "prompt_tokens": 2969,
      "reasoning_chunks": 0,
      "request_index": 27,
      "time_to_first_output_ms": 172.24420001730323,
      "ttft_ms": 172.24480002187192,
      "visible_generation_ms": 663.3137000026181
    },
    {
      "content_chunks": 81,
      "decode_tok_s": 114.63550386037498,
      "e2e_ms": 2005.406900017988,
      "effective_prefill_tok_s": 17111.040797546797,
      "generation_ms": 1831.893199996557,
      "mean_inter_token_latency_ms": 8.723300952364557,
      "output_token_source": "usage",
      "output_tokens": 211,
      "planned_context_tokens": 4096,
      "prompt_tokens": 2969,
      "reasoning_chunks": 0,
      "request_index": 28,
      "time_to_first_output_ms": 173.51370002143085,
      "ttft_ms": 173.5142000252381,
      "visible_generation_ms": 1831.8926999927498
    },
    {
      "content_chunks": 20,
      "decode_tok_s": 102.60095606757842,
      "e2e_ms": 648.8521000137553,
      "effective_prefill_tok_s": 15563.512917584903,
      "generation_ms": 458.085399994161,
      "mean_inter_token_latency_ms": 9.74649787221619,
      "output_token_source": "usage",
      "output_tokens": 48,
      "planned_context_tokens": 4096,
      "prompt_tokens": 2969,
      "reasoning_chunks": 0,
      "request_index": 29,
      "time_to_first_output_ms": 190.76670001959428,
      "ttft_ms": 190.76710002264008,
      "visible_generation_ms": 458.0849999911152
    },
    {
      "content_chunks": 17,
      "decode_tok_s": 97.39893577974007,
      "e2e_ms": 580.3942999918945,
      "effective_prefill_tok_s": 15606.085378124193,
      "generation_ms": 390.1480000349693,
      "mean_inter_token_latency_ms": 10.267052632499192,
      "output_token_source": "usage",
      "output_tokens": 39,
      "planned_context_tokens": 4096,
      "prompt_tokens": 2969,
      "reasoning_chunks": 0,
      "request_index": 30,
      "time_to_first_output_ms": 190.2462999569252,
      "ttft_ms": 190.2468999614939,
      "visible_generation_ms": 390.1474000304006
    },
    {
      "content_chunks": 12,
      "decode_tok_s": 83.28628864001001,
      "e2e_ms": 462.3188000405207,
      "effective_prefill_tok_s": 15948.397881211738,
      "generation_ms": 276.15590003551915,
      "mean_inter_token_latency_ms": 12.006778262413876,
      "output_token_source": "usage",
      "output_tokens": 24,
      "planned_context_tokens": 4096,
      "prompt_tokens": 2969,
      "reasoning_chunks": 0,
      "request_index": 31,
      "time_to_first_output_ms": 186.16290000500157,
      "ttft_ms": 186.16330000804737,
      "visible_generation_ms": 276.15550003247336
    },
    {
      "content_chunks": 69,
      "decode_tok_s": 105.26023146585864,
      "e2e_ms": 1798.1747000012547,
      "effective_prefill_tok_s": 16212.544324896444,
      "generation_ms": 1615.0448999833316,
      "mean_inter_token_latency_ms": 9.500264117549008,
      "output_token_source": "usage",
      "output_tokens": 171,
      "planned_context_tokens": 4096,
      "prompt_tokens": 2969,
      "reasoning_chunks": 0,
      "request_index": 32,
      "time_to_first_output_ms": 183.12980001792312,
      "ttft_ms": 183.1302000209689,
      "visible_generation_ms": 1615.0444999802858
    },
    {
      "content_chunks": 28,
      "decode_tok_s": 77.25105015450303,
      "e2e_ms": 852.8537999955006,
      "effective_prefill_tok_s": 15409.879974086922,
      "generation_ms": 660.1852000458166,
      "mean_inter_token_latency_ms": 12.94480784403562,
      "output_token_source": "usage",
      "output_tokens": 52,
      "planned_context_tokens": 4096,
      "prompt_tokens": 2969,
      "reasoning_chunks": 0,
      "request_index": 33,
      "time_to_first_output_ms": 192.668599949684,
      "ttft_ms": 192.66889995196834,
      "visible_generation_ms": 660.1849000435323
    },
    {
      "content_chunks": 33,
      "decode_tok_s": 101.92033193425594,
      "e2e_ms": 973.3081999584101,
      "effective_prefill_tok_s": 15760.579338730093,
      "generation_ms": 784.9267999990843,
      "mean_inter_token_latency_ms": 9.811584999988554,
      "output_token_source": "usage",
      "output_tokens": 81,
      "planned_context_tokens": 4096,
      "prompt_tokens": 2969,
      "reasoning_chunks": 0,
      "request_index": 34,
      "time_to_first_output_ms": 188.3813999593258,
      "ttft_ms": 188.3819999638945,
      "visible_generation_ms": 784.9261999945156
    },
    {
      "content_chunks": 50,
      "decode_tok_s": 96.37306597532844,
      "e2e_ms": 1320.5506000085734,
      "effective_prefill_tok_s": 15665.132336273304,
      "generation_ms": 1131.0213999822736,
      "mean_inter_token_latency_ms": 10.376343119103428,
      "output_token_source": "usage",
      "output_tokens": 110,
      "planned_context_tokens": 4096,
      "prompt_tokens": 2969,
      "reasoning_chunks": 0,
      "request_index": 35,
      "time_to_first_output_ms": 189.5292000262998,
      "ttft_ms": 189.5296000293456,
      "visible_generation_ms": 1131.0209999792278
    },
    {
      "content_chunks": 15,
      "decode_tok_s": 101.28284856532542,
      "e2e_ms": 532.6036000042222,
      "effective_prefill_tok_s": 15873.88998863284,
      "generation_ms": 345.5668999813497,
      "mean_inter_token_latency_ms": 9.873339999467134,
      "output_token_source": "usage",
      "output_tokens": 36,
      "planned_context_tokens": 4096,
      "prompt_tokens": 2969,
      "reasoning_chunks": 0,
      "request_index": 36,
      "time_to_first_output_ms": 187.03670002287254,
      "ttft_ms": 187.03710002591833,
      "visible_generation_ms": 345.5664999783039
    },
    {
      "content_chunks": 4,
      "decode_tok_s": 44.47963277252711,
      "e2e_ms": 275.24490002542734,
      "effective_prefill_tok_s": 16021.273919049838,
      "generation_ms": 89.92880000732839,
      "mean_inter_token_latency_ms": 22.482200001832098,
      "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": 185.31610001809895,
      "ttft_ms": 185.3164000203833,
      "visible_generation_ms": 89.92850000504404
    },
    {
      "content_chunks": 11,
      "decode_tok_s": 100.29512931600861,
      "e2e_ms": 401.06920001562685,
      "effective_prefill_tok_s": 17287.156614270407,
      "generation_ms": 229.32320000836626,
      "mean_inter_token_latency_ms": 9.970573913407229,
      "output_token_source": "usage",
      "output_tokens": 24,
      "planned_context_tokens": 4096,
      "prompt_tokens": 2969,
      "reasoning_chunks": 0,
      "request_index": 38,
      "time_to_first_output_ms": 171.7460000072606,
      "ttft_ms": 171.74630000954494,
      "visible_generation_ms": 229.3229000060819
    },
    {
      "content_chunks": 3,
      "decode_tok_s": 58.36663879058056,
      "e2e_ms": 251.3700000126846,
      "effective_prefill_tok_s": 16238.445352948153,
      "generation_ms": 68.5323000070639,
      "mean_inter_token_latency_ms": 17.133075001765974,
      "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": 182.83770000562072,
      "ttft_ms": 182.83810000866652,
      "visible_generation_ms": 68.5319000040181
    },
    {
      "content_chunks": 34,
      "decode_tok_s": 101.68631567716902,
      "e2e_ms": 944.7936999495141,
      "effective_prefill_tok_s": 15829.348009463267,
      "generation_ms": 757.23069999367,
      "mean_inter_token_latency_ms": 9.834164934982727,
      "output_token_source": "usage",
      "output_tokens": 78,
      "planned_context_tokens": 4096,
      "prompt_tokens": 2969,
      "reasoning_chunks": 0,
      "request_index": 40,
      "time_to_first_output_ms": 187.56299995584413,
      "ttft_ms": 187.56329995812848,
      "visible_generation_ms": 757.2303999913856
    },
    {
      "content_chunks": 27,
      "decode_tok_s": 98.92443984004427,
      "e2e_ms": 782.8876999556087,
      "effective_prefill_tok_s": 15921.884629534217,
      "generation_ms": 596.4147999766283,
      "mean_inter_token_latency_ms": 10.108725423332682,
      "output_token_source": "usage",
      "output_tokens": 60,
      "planned_context_tokens": 4096,
      "prompt_tokens": 2969,
      "reasoning_chunks": 0,
      "request_index": 41,
      "time_to_first_output_ms": 186.47289997898042,
      "ttft_ms": 186.47319998126477,
      "visible_generation_ms": 596.4144999743439
    },
    {
      "content_chunks": 15,
      "decode_tok_s": 102.06884266394081,
      "e2e_ms": 506.49169995449483,
      "effective_prefill_tok_s": 16208.05708061903,
      "generation_ms": 323.31119995797053,
      "mean_inter_token_latency_ms": 9.79730908963547,
      "output_token_source": "usage",
      "output_tokens": 34,
      "planned_context_tokens": 4096,
      "prompt_tokens": 2969,
      "reasoning_chunks": 0,
      "request_index": 42,
      "time_to_first_output_ms": 183.1804999965243,
      "ttft_ms": 183.18079999880865,
      "visible_generation_ms": 323.3108999556862
    },
    {
      "content_chunks": 12,
      "decode_tok_s": 98.44310263810837,
      "e2e_ms": 435.06849999539554,
      "effective_prefill_tok_s": 16392.926693888956,
      "generation_ms": 253.9538000128232,
      "mean_inter_token_latency_ms": 10.158152000512928,
      "output_token_source": "usage",
      "output_tokens": 26,
      "planned_context_tokens": 4096,
      "prompt_tokens": 2969,
      "reasoning_chunks": 0,
      "request_index": 43,
      "time_to_first_output_ms": 181.11469998257235,
      "ttft_ms": 181.1149999848567,
      "visible_generation_ms": 253.95350001053885
    },
    {
      "content_chunks": 15,
      "decode_tok_s": 89.46652841439992,
      "e2e_ms": 531.0260999831371,
      "effective_prefill_tok_s": 16089.716562075137,
      "generation_ms": 346.4982999721542,
      "mean_inter_token_latency_ms": 11.177364515230781,
      "output_token_source": "usage",
      "output_tokens": 32,
      "planned_context_tokens": 4096,
      "prompt_tokens": 2969,
      "reasoning_chunks": 0,
      "request_index": 44,
      "time_to_first_output_ms": 184.5278000109829,
      "ttft_ms": 184.5282000140287,
      "visible_generation_ms": 346.4978999691084
    },
    {
      "content_chunks": 16,
      "decode_tok_s": 111.65169897797172,
      "e2e_ms": 550.1232000533491,
      "effective_prefill_tok_s": 16232.044425726534,
      "generation_ms": 367.2134000225924,
      "mean_inter_token_latency_ms": 8.956424390794936,
      "output_token_source": "usage",
      "output_tokens": 42,
      "planned_context_tokens": 4096,
      "prompt_tokens": 2969,
      "reasoning_chunks": 0,
      "request_index": 45,
      "time_to_first_output_ms": 182.90980003075674,
      "ttft_ms": 182.9101000330411,
      "visible_generation_ms": 367.21310002030805
    },
    {
      "content_chunks": 13,
      "decode_tok_s": 120.56726899704691,
      "e2e_ms": 481.39179998543113,
      "effective_prefill_tok_s": 16241.501112843518,
      "generation_ms": 298.58850000891834,
      "mean_inter_token_latency_ms": 8.294125000247732,
      "output_token_source": "usage",
      "output_tokens": 37,
      "planned_context_tokens": 4096,
      "prompt_tokens": 2969,
      "reasoning_chunks": 0,
      "request_index": 46,
      "time_to_first_output_ms": 182.8032999765128,
      "ttft_ms": 182.8036999795586,
      "visible_generation_ms": 298.58810000587255
    },
    {
      "content_chunks": 41,
      "decode_tok_s": 102.46532654162681,
      "e2e_ms": 1111.8476000265218,
      "effective_prefill_tok_s": 16074.306716024028,
      "generation_ms": 927.1429000073113,
      "mean_inter_token_latency_ms": 9.759398947445382,
      "output_token_source": "usage",
      "output_tokens": 96,
      "planned_context_tokens": 4096,
      "prompt_tokens": 2969,
      "reasoning_chunks": 0,
      "request_index": 47,
      "time_to_first_output_ms": 184.70470001921058,
      "ttft_ms": 184.70520002301782,
      "visible_generation_ms": 927.142400003504
    },
    {
      "content_chunks": 3,
      "decode_tok_s": 60.031216256998555,
      "e2e_ms": 236.74669995671138,
      "effective_prefill_tok_s": 17452.930289268254,
      "generation_ms": 66.63199997274205,
      "mean_inter_token_latency_ms": 16.657999993185513,
      "output_token_source": "usage",
      "output_tokens": 5,
      "planned_context_tokens": 4096,
      "prompt_tokens": 2969,
      "reasoning_chunks": 0,
      "request_index": 48,
      "time_to_first_output_ms": 170.11469998396933,
      "ttft_ms": 170.11509998701513,
      "visible_generation_ms": 66.63159996969625
    },
    {
      "content_chunks": 13,
      "decode_tok_s": 95.99911253477991,
      "e2e_ms": 464.6729999803938,
      "effective_prefill_tok_s": 16186.858171925398,
      "generation_ms": 281.2526000197977,
      "mean_inter_token_latency_ms": 10.416762963696211,
      "output_token_source": "usage",
      "output_tokens": 28,
      "planned_context_tokens": 4096,
      "prompt_tokens": 2969,
      "reasoning_chunks": 0,
      "request_index": 49,
      "time_to_first_output_ms": 183.42039996059611,
      "ttft_ms": 183.42089996440336,
      "visible_generation_ms": 281.25210001599044
    },
    {
      "content_chunks": 95,
      "decode_tok_s": 111.78765037735374,
      "e2e_ms": 2363.832500006538,
      "effective_prefill_tok_s": 16392.202635237234,
      "generation_ms": 2182.7098000212573,
      "mean_inter_token_latency_ms": 8.945531967300235,
      "output_token_source": "usage",
      "output_tokens": 245,
      "planned_context_tokens": 4096,
      "prompt_tokens": 2969,
      "reasoning_chunks": 0,
      "request_index": 50,
      "time_to_first_output_ms": 181.12269998528063,
      "ttft_ms": 181.12299998756498,
      "visible_generation_ms": 2182.709500018973
    },
    {
      "content_chunks": 95,
      "decode_tok_s": 113.18145036184767,
      "e2e_ms": 2426.5616000047885,
      "effective_prefill_tok_s": 16279.411507153634,
      "generation_ms": 2244.1840000101365,
      "mean_inter_token_latency_ms": 8.835370078780064,
      "output_token_source": "usage",
      "output_tokens": 255,
      "planned_context_tokens": 4096,
      "prompt_tokens": 2969,
      "reasoning_chunks": 0,
      "request_index": 51,
      "time_to_first_output_ms": 182.37759999465197,
      "ttft_ms": 182.37799999769777,
      "visible_generation_ms": 2244.1836000070907
    },
    {
      "content_chunks": 39,
      "decode_tok_s": 107.15876402938835,
      "e2e_ms": 1058.3332999958657,
      "effective_prefill_tok_s": 16391.52388866841,
      "generation_ms": 877.203100011684,
      "mean_inter_token_latency_ms": 9.331947872464724,
      "output_token_source": "usage",
      "output_tokens": 95,
      "planned_context_tokens": 4096,
      "prompt_tokens": 2969,
      "reasoning_chunks": 0,
      "request_index": 52,
      "time_to_first_output_ms": 181.13019998418167,
      "ttft_ms": 181.13049998646602,
      "visible_generation_ms": 877.2028000093997
    },
    {
      "content_chunks": 3,
      "decode_tok_s": 61.50326275324345,
      "e2e_ms": 239.78839995106682,
      "effective_prefill_tok_s": 16989.869029447345,
      "generation_ms": 65.03719999454916,
      "mean_inter_token_latency_ms": 16.25929999863729,
      "output_token_source": "usage",
      "output_tokens": 5,
      "planned_context_tokens": 4096,
      "prompt_tokens": 2969,
      "reasoning_chunks": 0,
      "request_index": 53,
      "time_to_first_output_ms": 174.75119995651767,
      "ttft_ms": 174.75159995956346,
      "visible_generation_ms": 65.03679999150336
    },
    {
      "content_chunks": 74,
      "decode_tok_s": 116.42820837083985,
      "e2e_ms": 1874.8355999705382,
      "effective_prefill_tok_s": 17041.984467960952,
      "generation_ms": 1700.6187999504618,
      "mean_inter_token_latency_ms": 8.588983838133645,
      "output_token_source": "usage",
      "output_tokens": 199,
      "planned_context_tokens": 4096,
      "prompt_tokens": 2969,
      "reasoning_chunks": 0,
      "request_index": 54,
      "time_to_first_output_ms": 174.21680002007633,
      "ttft_ms": 174.21720002312213,
      "visible_generation_ms": 1700.618399947416
    },
    {
      "content_chunks": 11,
      "decode_tok_s": 92.15384601119753,
      "e2e_ms": 411.980000033509,
      "effective_prefill_tok_s": 16127.0873131244,
      "generation_ms": 227.87980001885444,
      "mean_inter_token_latency_ms": 10.851419048516878,
      "output_token_source": "usage",
      "output_tokens": 22,
      "planned_context_tokens": 4096,
      "prompt_tokens": 2969,
      "reasoning_chunks": 0,
      "request_index": 55,
      "time_to_first_output_ms": 184.10020001465455,
      "ttft_ms": 184.10060001770034,
      "visible_generation_ms": 227.87940001580864
    },
    {
      "content_chunks": 14,
      "decode_tok_s": 106.65577761242322,
      "e2e_ms": 516.8641000054777,
      "effective_prefill_tok_s": 15733.50234579135,
      "generation_ms": 328.1585000222549,
      "mean_inter_token_latency_ms": 9.375957143492997,
      "output_token_source": "usage",
      "output_tokens": 36,
      "planned_context_tokens": 4096,
      "prompt_tokens": 2969,
      "reasoning_chunks": 0,
      "request_index": 56,
      "time_to_first_output_ms": 188.70559998322278,
      "ttft_ms": 188.70589998550713,
      "visible_generation_ms": 328.15820001997054
    },
    {
      "content_chunks": 40,
      "decode_tok_s": 119.27914077411654,
      "e2e_ms": 1123.3421000069939,
      "effective_prefill_tok_s": 16103.644770133878,
      "generation_ms": 938.9738999889232,
      "mean_inter_token_latency_ms": 8.383695535615386,
      "output_token_source": "usage",
      "output_tokens": 113,
      "planned_context_tokens": 4096,
      "prompt_tokens": 2969,
      "reasoning_chunks": 0,
      "request_index": 57,
      "time_to_first_output_ms": 184.36820001807064,
      "ttft_ms": 184.36880002263933,
      "visible_generation_ms": 938.9732999843545
    },
    {
      "content_chunks": 45,
      "decode_tok_s": 116.14204376290724,
      "e2e_ms": 1265.7783000031486,
      "effective_prefill_tok_s": 15666.719434618593,
      "generation_ms": 1076.2683000066318,
      "mean_inter_token_latency_ms": 8.610146400053054,
      "output_token_source": "usage",
      "output_tokens": 126,
      "planned_context_tokens": 4096,
      "prompt_tokens": 2969,
      "reasoning_chunks": 0,
      "request_index": 58,
      "time_to_first_output_ms": 189.50999999651685,
      "ttft_ms": 189.5102999988012,
      "visible_generation_ms": 1076.2680000043474
    },
    {
      "content_chunks": 45,
      "decode_tok_s": 113.65024157579539,
      "e2e_ms": 1220.308899995871,
      "effective_prefill_tok_s": 16309.987545449001,
      "generation_ms": 1038.2731999852695,
      "mean_inter_token_latency_ms": 8.79892542360398,
      "output_token_source": "usage",
      "output_tokens": 119,
      "planned_context_tokens": 4096,
      "prompt_tokens": 2969,
      "reasoning_chunks": 0,
      "request_index": 59,
      "time_to_first_output_ms": 182.03570001060143,
      "ttft_ms": 182.03610001364723,
      "visible_generation_ms": 1038.2727999822237
    }
  ],
  "requests": 60,
  "run_id": "benchmark-20260902T112047Z",
  "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": "2026-09-02T11:20:47Z",
  "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": 56269.93650000077
}
