{ "base_url": "http://127.0.0.1:39272/v1", "completed": 100, "concurrency": 8, "context_distribution": { "4096": 100 }, "context_seed": 0, "context_tokens": 4096, "engine": "sglang", "failed": 0, "failures": [], "finished_at": "2026-09-05T03:59:32Z", "gpu": "Primary Node / RTX PRO 6000 Blackwell Max-Q / TP=1", "manifest": null, "max_context_tokens": 262144, "max_tokens": 512, "measurement_protocol": "capacity-v3", "metrics": { "completion_tokens_max": 512.0, "completion_tokens_mean": 512.0, "completion_tokens_mean_ci95_high": 512.0, "completion_tokens_mean_ci95_low": 512.0, "completion_tokens_min": 512.0, "completion_tokens_p25": 512.0, "completion_tokens_p50": 512.0, "completion_tokens_p75": 512.0, "completion_tokens_p90": 512.0, "completion_tokens_p95": 512.0, "completion_tokens_p99": 512.0, "completion_tokens_samples": 100, "completion_tokens_stddev": 0.0, "content_chunks": 4453, "content_chunks_s": null, "decode_tok_s_max": 362.26567046056687, "decode_tok_s_mean": 201.47008045953334, "decode_tok_s_mean_ci95_high": 214.18631627931757, "decode_tok_s_mean_ci95_low": 188.7538446397491, "decode_tok_s_min": 119.59990904850136, "decode_tok_s_p25": 139.3695912457905, "decode_tok_s_p50": 182.48562925045633, "decode_tok_s_p75": 255.31331490930023, "decode_tok_s_p90": 297.6178168974192, "decode_tok_s_p95": 298.35353222805566, "decode_tok_s_p99": 305.9929528291385, "decode_tok_s_samples": 100, "decode_tok_s_stddev": 64.87875418257268, "e2e_max_ms": 9785.917600034736, "e2e_mean_ci95_high_ms": 5576.569880195789, "e2e_mean_ci95_low_ms": 4869.775121801172, "e2e_mean_ms": 5223.17250099848, "e2e_min_ms": 2016.2611999548972, "e2e_p25_ms": 4825.420800014399, "e2e_p50_ms": 4856.431000051089, "e2e_p75_ms": 4938.591299927793, "e2e_p90_ms": 9421.03229998611, "e2e_p95_ms": 9578.948900103569, "e2e_p99_ms": 9758.101999992505, "e2e_samples": 100, "e2e_stddev_ms": 1803.0478530474913, "effective_prefill_tok_s_max": 12439.916748860187, "effective_prefill_tok_s_mean": 2609.204376255017, "effective_prefill_tok_s_mean_ci95_high": 3184.4903140957817, "effective_prefill_tok_s_mean_ci95_low": 2033.9184384142523, "effective_prefill_tok_s_min": 651.5550420116475, "effective_prefill_tok_s_p25": 1219.3472930148669, "effective_prefill_tok_s_p50": 1515.3047852290294, "effective_prefill_tok_s_p75": 2606.3988967970195, "effective_prefill_tok_s_p90": 3661.888374489671, "effective_prefill_tok_s_p95": 12009.050779381272, "effective_prefill_tok_s_p99": 12072.146242093722, "effective_prefill_tok_s_samples": 100, "effective_prefill_tok_s_stddev": 2935.132335922269, "generation_max_ms": 4272.578499978408, "generation_mean_ci95_high_ms": 2979.3411553751644, "generation_mean_ci95_low_ms": 2635.5788166184943, "generation_mean_ms": 2807.4599859968293, "generation_min_ms": 1410.5670000426471, "generation_p25_ms": 1944.9200999224558, "generation_p50_ms": 2774.777099955827, "generation_p75_ms": 3649.7414000332355, "generation_p90_ms": 4147.189799929038, "generation_p95_ms": 4175.375899998471, "generation_p99_ms": 4256.237900000997, "generation_samples": 100, "generation_stddev_ms": 876.9447417261993, "mean_inter_token_latency_ms_max": 8.361210371777707, "mean_inter_token_latency_ms_mean": 5.494050853222758, "mean_inter_token_latency_ms_mean_ci95_high": 5.830413219912259, "mean_inter_token_latency_ms_mean_ci95_low": 5.157688486533257, "mean_inter_token_latency_ms_min": 2.7604050881460807, "mean_inter_token_latency_ms_p25": 3.8061058706897373, "mean_inter_token_latency_ms_p50": 5.4300921721249065, "mean_inter_token_latency_ms_p75": 7.14235107638598, "mean_inter_token_latency_ms_p90": 8.115831311015729, "mean_inter_token_latency_ms_p95": 8.17099001956648, "mean_inter_token_latency_ms_p99": 8.329232681019564, "mean_inter_token_latency_ms_samples": 100, "mean_inter_token_latency_ms_stddev": 1.716134523926026, "output_token_sources": { "usage": 100 }, "output_tokens": 51200, "prefix_cache_hit_avg": null, "prompt_token_samples": 100, "prompt_tokens": 365425, "prompt_tokens_max": 3689.0, "prompt_tokens_mean": 3654.25, "prompt_tokens_mean_ci95_high": 3658.3042454864803, "prompt_tokens_mean_ci95_low": 3650.1957545135197, "prompt_tokens_min": 3613.0, "prompt_tokens_p25": 3636.0, "prompt_tokens_p50": 3655.0, "prompt_tokens_p75": 3672.0, "prompt_tokens_p90": 3680.0, "prompt_tokens_p95": 3685.0, "prompt_tokens_p99": 3689.0, "prompt_tokens_samples": 100, "prompt_tokens_stddev": 20.68492595143047, "reasoning_chunks": 0, "request_canary_passed": 100, "request_canary_samples": 100, "throughput_tok_s": 764.2997714409191, "time_to_first_output_max_ms": 5585.099900024943, "time_to_first_output_mean_ci95_high_ms": 2692.2523358979893, "time_to_first_output_mean_ci95_low_ms": 2139.1726941053125, "time_to_first_output_mean_ms": 2415.712515001651, "time_to_first_output_min_ms": 290.596800041385, "time_to_first_output_p25_ms": 1398.643899941817, "time_to_first_output_p50_ms": 2331.260900013149, "time_to_first_output_p75_ms": 2995.8651000633836, "time_to_first_output_p90_ms": 5295.582200051285, "time_to_first_output_p95_ms": 5339.059900026768, "time_to_first_output_p99_ms": 5513.339100056328, "time_to_first_output_samples": 100, "time_to_first_output_stddev_ms": 1410.9174535527468, "tpot_max_ms": 8.361210371777707, "tpot_mean_ci95_high_ms": 5.830413219912259, "tpot_mean_ci95_low_ms": 5.157688486533257, "tpot_mean_ms": 5.494050853222758, "tpot_min_ms": 2.7604050881460807, "tpot_p25_ms": 3.8061058706897373, "tpot_p50_ms": 5.4300921721249065, "tpot_p75_ms": 7.14235107638598, "tpot_p90_ms": 8.115831311015729, "tpot_p95_ms": 8.17099001956648, "tpot_p99_ms": 8.329232681019564, "tpot_samples": 100, "tpot_stddev_ms": 1.716134523926026, "ttft_max_ms": 5585.100500029512, "ttft_mean_ci95_high_ms": 2692.25281702049, "ttft_mean_ci95_low_ms": 2139.1731609830454, "ttft_mean_ms": 2415.7129890017677, "ttft_min_ms": 290.5974000459537, "ttft_p25_ms": 1398.6444999463856, "ttft_p50_ms": 2331.2613000161946, "ttft_p75_ms": 2995.865600067191, "ttft_p90_ms": 5295.582700055093, "ttft_p95_ms": 5339.060400030576, "ttft_p99_ms": 5513.339600060135, "ttft_samples": 100, "ttft_stddev_ms": 1410.9174898914403 }, "model": "qwen38-27b-inferact-nvfp4-pro6000-dflash2-k12-chunk1k", "prompt_cache_mode": "unique", "request_canaries": true, "request_timings": [ { "content_chunks": 44, "decode_tok_s": 297.8928135078324, "e2e_ms": 2017.7082000300288, "effective_prefill_tok_s": 11950.671805183245, "generation_ms": 1715.3820999665186, "mean_inter_token_latency_ms": 3.356912133006886, "output_token_source": "usage", "output_tokens": 512, "planned_context_tokens": 4096, "prompt_tokens": 3613, "reasoning_chunks": 0, "request_canary": { "foreign_marker_count": 0, "marker": "ANVIL_REQ_0_00000", "passed": true }, "request_index": 0, "time_to_first_output_ms": 302.32610006351024, "tpot_ms": 3.356912133006886, "ttft_ms": 302.32630006503314, "visible_content_truncated": false, "visible_generation_ms": 1715.3818999649957 }, { "content_chunks": 44, "decode_tok_s": 295.8586219570765, "e2e_ms": 2017.7731000585482, "effective_prefill_tok_s": 12439.916748860187, "generation_ms": 1727.1763000171632, "mean_inter_token_latency_ms": 3.3799927593290864, "output_token_source": "usage", "output_tokens": 512, "planned_context_tokens": 4096, "prompt_tokens": 3615, "reasoning_chunks": 0, "request_canary": { "foreign_marker_count": 0, "marker": "ANVIL_REQ_0_00001", "passed": true }, "request_index": 1, "time_to_first_output_ms": 290.596800041385, "tpot_ms": 3.3799927593290864, "ttft_ms": 290.5974000459537, "visible_content_truncated": false, "visible_generation_ms": 1727.1757000125945 }, { "content_chunks": 44, "decode_tok_s": 297.62078103077994, "e2e_ms": 2018.22099997662, "effective_prefill_tok_s": 12005.802087329723, "generation_ms": 1716.95000003092, "mean_inter_token_latency_ms": 3.3599804305888843, "output_token_source": "usage", "output_tokens": 512, "planned_context_tokens": 4096, "prompt_tokens": 3617, "reasoning_chunks": 0, "request_canary": { "foreign_marker_count": 0, "marker": "ANVIL_REQ_0_00002", "passed": true }, "request_index": 2, "time_to_first_output_ms": 301.27099994570017, "tpot_ms": 3.3599804305888843, "ttft_ms": 301.27159995026886, "visible_content_truncated": false, "visible_generation_ms": 1716.9494000263512 }, { "content_chunks": 44, "decode_tok_s": 297.6178168974192, "e2e_ms": 2018.1566000683233, "effective_prefill_tok_s": 12009.050779381272, "generation_ms": 1716.9671000447124, "mean_inter_token_latency_ms": 3.360013894412353, "output_token_source": "usage", "output_tokens": 512, "planned_context_tokens": 4096, "prompt_tokens": 3617, "reasoning_chunks": 0, "request_canary": { "foreign_marker_count": 0, "marker": "ANVIL_REQ_0_00003", "passed": true }, "request_index": 3, "time_to_first_output_ms": 301.1895000236109, "tpot_ms": 3.360013894412353, "ttft_ms": 301.1897000251338, "visible_content_truncated": false, "visible_generation_ms": 1716.9669000431895 }, { "content_chunks": 44, "decode_tok_s": 297.73190051524176, "e2e_ms": 2017.0812000287697, "effective_prefill_tok_s": 12032.370034933812, "generation_ms": 1716.309200040996, "mean_inter_token_latency_ms": 3.3587264188669197, "output_token_source": "usage", "output_tokens": 512, "planned_context_tokens": 4096, "prompt_tokens": 3619, "reasoning_chunks": 0, "request_canary": { "foreign_marker_count": 0, "marker": "ANVIL_REQ_0_00004", "passed": true }, "request_index": 4, "time_to_first_output_ms": 300.7719999877736, "tpot_ms": 3.3587264188669197, "ttft_ms": 300.7721999892965, "visible_content_truncated": false, "visible_generation_ms": 1716.3090000394732 }, { "content_chunks": 45, "decode_tok_s": 290.9510963656616, "e2e_ms": 2056.9170999806374, "effective_prefill_tok_s": 12042.253004894914, "generation_ms": 1756.3088999595493, "mean_inter_token_latency_ms": 3.4370037181204487, "output_token_source": "usage", "output_tokens": 512, "planned_context_tokens": 4096, "prompt_tokens": 3620, "reasoning_chunks": 0, "request_canary": { "foreign_marker_count": 0, "marker": "ANVIL_REQ_0_00005", "passed": true }, "request_index": 5, "time_to_first_output_ms": 300.6082000210881, "tpot_ms": 3.4370037181204487, "ttft_ms": 300.6083000218496, "visible_content_truncated": false, "visible_generation_ms": 1756.3087999587879 }, { "content_chunks": 44, "decode_tok_s": 297.74534521777537, "e2e_ms": 2016.2611999548972, "effective_prefill_tok_s": 12072.146242093722, "generation_ms": 1716.2317000329494, "mean_inter_token_latency_ms": 3.358574755446085, "output_token_source": "usage", "output_tokens": 512, "planned_context_tokens": 4096, "prompt_tokens": 3622, "reasoning_chunks": 0, "request_canary": { "foreign_marker_count": 0, "marker": "ANVIL_REQ_0_00006", "passed": true }, "request_index": 6, "time_to_first_output_ms": 300.0294999219477, "tpot_ms": 3.358574755446085, "ttft_ms": 300.0296999234706, "visible_content_truncated": false, "visible_generation_ms": 1716.2315000314265 }, { "content_chunks": 45, "decode_tok_s": 290.9999910148762, "e2e_ms": 2056.1076999874786, "effective_prefill_tok_s": 12069.555559905453, "generation_ms": 1756.0137999244034, "mean_inter_token_latency_ms": 3.436426222944038, "output_token_source": "usage", "output_tokens": 512, "planned_context_tokens": 4096, "prompt_tokens": 3622, "reasoning_chunks": 0, "request_canary": { "foreign_marker_count": 0, "marker": "ANVIL_REQ_0_00007", "passed": true }, "request_index": 7, "time_to_first_output_ms": 300.0939000630751, "tpot_ms": 3.436426222944038, "ttft_ms": 300.094100064598, "visible_content_truncated": false, "visible_generation_ms": 1756.0135999228805 }, { "content_chunks": 44, "decode_tok_s": 122.84678258527472, "e2e_ms": 4831.750100012869, "effective_prefill_tok_s": 5389.103060530554, "generation_ms": 4159.653099952266, "mean_inter_token_latency_ms": 8.140221330630657, "output_token_source": "usage", "output_tokens": 512, "planned_context_tokens": 4096, "prompt_tokens": 3622, "reasoning_chunks": 0, "request_canary": { "foreign_marker_count": 0, "marker": "ANVIL_REQ_0_00008", "passed": true }, "request_index": 8, "time_to_first_output_ms": 672.097000060603, "tpot_ms": 8.140221330630657, "ttft_ms": 672.0975000644103, "visible_content_truncated": false, "visible_generation_ms": 4159.652599948458 }, { "content_chunks": 44, "decode_tok_s": 210.31229935832974, "e2e_ms": 4830.5873000063, "effective_prefill_tok_s": 1509.4545208777633, "generation_ms": 2429.7200000146404, "mean_inter_token_latency_ms": 4.754833659519845, "output_token_source": "usage", "output_tokens": 512, "planned_context_tokens": 4096, "prompt_tokens": 3624, "reasoning_chunks": 0, "request_canary": { "foreign_marker_count": 0, "marker": "ANVIL_REQ_0_00009", "passed": true }, "request_index": 9, "time_to_first_output_ms": 2400.86729999166, "tpot_ms": 4.754833659519845, "ttft_ms": 2400.867799995467, "visible_content_truncated": false, "visible_generation_ms": 2429.719500010833 }, { "content_chunks": 44, "decode_tok_s": 182.48562925045633, "e2e_ms": 4829.901200020686, "effective_prefill_tok_s": 1786.4878944427549, "generation_ms": 2800.2205000957474, "mean_inter_token_latency_ms": 5.479883561831208, "output_token_source": "usage", "output_tokens": 512, "planned_context_tokens": 4096, "prompt_tokens": 3626, "reasoning_chunks": 0, "request_canary": { "foreign_marker_count": 0, "marker": "ANVIL_REQ_0_00010", "passed": true }, "request_index": 10, "time_to_first_output_ms": 2029.6806999249384, "tpot_ms": 5.479883561831208, "ttft_ms": 2029.6813999302685, "visible_content_truncated": false, "visible_generation_ms": 2800.2198000904173 }, { "content_chunks": 44, "decode_tok_s": 155.0762448572613, "e2e_ms": 4830.595400067978, "effective_prefill_tok_s": 2362.1861090990406, "generation_ms": 3295.153300045058, "mean_inter_token_latency_ms": 6.448440900283871, "output_token_source": "usage", "output_tokens": 512, "planned_context_tokens": 4096, "prompt_tokens": 3627, "reasoning_chunks": 0, "request_canary": { "foreign_marker_count": 0, "marker": "ANVIL_REQ_0_00011", "passed": true }, "request_index": 11, "time_to_first_output_ms": 1535.4421000229195, "tpot_ms": 6.448440900283871, "ttft_ms": 1535.4427000274882, "visible_content_truncated": false, "visible_generation_ms": 3295.1527000404894 }, { "content_chunks": 44, "decode_tok_s": 139.3695912457905, "e2e_ms": 4831.039299955592, "effective_prefill_tok_s": 3114.563111610437, "generation_ms": 3666.510000010021, "mean_inter_token_latency_ms": 7.175166340528417, "output_token_source": "usage", "output_tokens": 512, "planned_context_tokens": 4096, "prompt_tokens": 3627, "reasoning_chunks": 0, "request_canary": { "foreign_marker_count": 0, "marker": "ANVIL_REQ_0_00012", "passed": true }, "request_index": 12, "time_to_first_output_ms": 1164.5292999455705, "tpot_ms": 7.175166340528417, "ttft_ms": 1164.5297999493778, "visible_content_truncated": false, "visible_generation_ms": 3666.509500006214 }, { "content_chunks": 44, "decode_tok_s": 244.1866143898795, "e2e_ms": 4830.390300019644, "effective_prefill_tok_s": 1325.1861899186365, "generation_ms": 2092.661799979396, "mean_inter_token_latency_ms": 4.095228571388251, "output_token_source": "usage", "output_tokens": 512, "planned_context_tokens": 4096, "prompt_tokens": 3628, "reasoning_chunks": 0, "request_canary": { "foreign_marker_count": 0, "marker": "ANVIL_REQ_0_00013", "passed": true }, "request_index": 13, "time_to_first_output_ms": 2737.728500040248, "tpot_ms": 4.095228571388251, "ttft_ms": 2737.728900043294, "visible_content_truncated": false, "visible_generation_ms": 2092.6613999763504 }, { "content_chunks": 44, "decode_tok_s": 123.2159666308843, "e2e_ms": 9610.0508000236, "effective_prefill_tok_s": 664.3039242508974, "generation_ms": 4147.189799929038, "mean_inter_token_latency_ms": 8.115831311015729, "output_token_source": "usage", "output_tokens": 512, "planned_context_tokens": 4096, "prompt_tokens": 3629, "reasoning_chunks": 0, "request_canary": { "foreign_marker_count": 0, "marker": "ANVIL_REQ_0_00014", "passed": true }, "request_index": 14, "time_to_first_output_ms": 5462.861000094563, "tpot_ms": 8.115831311015729, "ttft_ms": 5462.861599982716, "visible_content_truncated": false, "visible_generation_ms": 4147.189200040884 }, { "content_chunks": 45, "decode_tok_s": 287.1958372727522, "e2e_ms": 4966.19509998709, "effective_prefill_tok_s": 1139.030287969906, "generation_ms": 1779.2736999690533, "mean_inter_token_latency_ms": 3.481944618334742, "output_token_source": "usage", "output_tokens": 512, "planned_context_tokens": 4096, "prompt_tokens": 3630, "reasoning_chunks": 0, "request_canary": { "foreign_marker_count": 0, "marker": "ANVIL_REQ_0_00015", "passed": true }, "request_index": 15, "time_to_first_output_ms": 3186.9214000180364, "tpot_ms": 3.481944618334742, "ttft_ms": 3186.922000022605, "visible_content_truncated": false, "visible_generation_ms": 1779.2730999644846 }, { "content_chunks": 44, "decode_tok_s": 155.92979179929728, "e2e_ms": 4819.957100087777, "effective_prefill_tok_s": 2352.1539351923916, "generation_ms": 3277.115900069475, "mean_inter_token_latency_ms": 6.4131426615841, "output_token_source": "usage", "output_tokens": 512, "planned_context_tokens": 4096, "prompt_tokens": 3629, "reasoning_chunks": 0, "request_canary": { "foreign_marker_count": 0, "marker": "ANVIL_REQ_0_00016", "passed": true }, "request_index": 16, "time_to_first_output_ms": 1542.8412000183016, "tpot_ms": 6.4131426615841, "ttft_ms": 1542.8416000213474, "visible_content_truncated": false, "visible_generation_ms": 3277.1155000664294 }, { "content_chunks": 44, "decode_tok_s": 140.00991960563198, "e2e_ms": 4819.132099975832, "effective_prefill_tok_s": 3104.18066449322, "generation_ms": 3649.7414000332355, "mean_inter_token_latency_ms": 7.14235107638598, "output_token_source": "usage", "output_tokens": 512, "planned_context_tokens": 4096, "prompt_tokens": 3630, "reasoning_chunks": 0, "request_canary": { "foreign_marker_count": 0, "marker": "ANVIL_REQ_0_00017", "passed": true }, "request_index": 17, "time_to_first_output_ms": 1169.3906999425963, "tpot_ms": 7.14235107638598, "ttft_ms": 1169.391099945642, "visible_content_truncated": false, "visible_generation_ms": 3649.7410000301898 }, { "content_chunks": 45, "decode_tok_s": 262.7357288458141, "e2e_ms": 4857.007799902931, "effective_prefill_tok_s": 1246.8717889314753, "generation_ms": 1944.9200999224558, "mean_inter_token_latency_ms": 3.8061058706897373, "output_token_source": "usage", "output_tokens": 512, "planned_context_tokens": 4096, "prompt_tokens": 3631, "reasoning_chunks": 0, "request_canary": { "foreign_marker_count": 0, "marker": "ANVIL_REQ_0_00018", "passed": true }, "request_index": 18, "time_to_first_output_ms": 2912.087699980475, "tpot_ms": 3.8061058706897373, "ttft_ms": 2912.088099983521, "visible_content_truncated": false, "visible_generation_ms": 1944.91969991941 }, { "content_chunks": 45, "decode_tok_s": 181.39055704479975, "e2e_ms": 4856.431000051089, "effective_prefill_tok_s": 1780.998656760545, "generation_ms": 2817.125700064935, "mean_inter_token_latency_ms": 5.5129661449411635, "output_token_source": "usage", "output_tokens": 512, "planned_context_tokens": 4096, "prompt_tokens": 3632, "reasoning_chunks": 0, "request_canary": { "foreign_marker_count": 0, "marker": "ANVIL_REQ_0_00019", "passed": true }, "request_index": 19, "time_to_first_output_ms": 2039.3052999861538, "tpot_ms": 5.5129661449411635, "ttft_ms": 2039.305799989961, "visible_content_truncated": false, "visible_generation_ms": 2817.1252000611275 }, { "content_chunks": 45, "decode_tok_s": 209.08069794549306, "e2e_ms": 4856.727699982002, "effective_prefill_tok_s": 1505.78483742835, "generation_ms": 2444.032400031574, "mean_inter_token_latency_ms": 4.782842270120497, "output_token_source": "usage", "output_tokens": 512, "planned_context_tokens": 4096, "prompt_tokens": 3633, "reasoning_chunks": 0, "request_canary": { "foreign_marker_count": 0, "marker": "ANVIL_REQ_0_00020", "passed": true }, "request_index": 20, "time_to_first_output_ms": 2412.6952999504283, "tpot_ms": 4.782842270120497, "ttft_ms": 2412.695699953474, "visible_content_truncated": false, "visible_generation_ms": 2444.0320000285283 }, { "content_chunks": 45, "decode_tok_s": 302.51081609536124, "e2e_ms": 4856.225499999709, "effective_prefill_tok_s": 1147.447401579309, "generation_ms": 1689.1957999905571, "mean_inter_token_latency_ms": 3.305666927574476, "output_token_source": "usage", "output_tokens": 512, "planned_context_tokens": 4096, "prompt_tokens": 3634, "reasoning_chunks": 0, "request_canary": { "foreign_marker_count": 0, "marker": "ANVIL_REQ_0_00021", "passed": true }, "request_index": 21, "time_to_first_output_ms": 3167.0297000091523, "tpot_ms": 3.305666927574476, "ttft_ms": 3167.030300013721, "visible_content_truncated": false, "visible_generation_ms": 1689.1951999859884 }, { "content_chunks": 45, "decode_tok_s": 120.05907846454737, "e2e_ms": 9581.485099974088, "effective_prefill_tok_s": 682.5974200819011, "generation_ms": 4256.237900000997, "mean_inter_token_latency_ms": 8.329232681019564, "output_token_source": "usage", "output_tokens": 512, "planned_context_tokens": 4096, "prompt_tokens": 3635, "reasoning_chunks": 0, "request_canary": { "foreign_marker_count": 0, "marker": "ANVIL_REQ_0_00022", "passed": true }, "request_index": 22, "time_to_first_output_ms": 5325.2471999730915, "tpot_ms": 8.329232681019564, "ttft_ms": 5325.24779997766, "visible_content_truncated": false, "visible_generation_ms": 4256.237299996428 }, { "content_chunks": 45, "decode_tok_s": 136.1072614704381, "e2e_ms": 4937.867900007404, "effective_prefill_tok_s": 3072.305648844737, "generation_ms": 3754.391900030896, "mean_inter_token_latency_ms": 7.3471465754029275, "output_token_source": "usage", "output_tokens": 512, "planned_context_tokens": 4096, "prompt_tokens": 3636, "reasoning_chunks": 0, "request_canary": { "foreign_marker_count": 0, "marker": "ANVIL_REQ_0_00023", "passed": true }, "request_index": 23, "time_to_first_output_ms": 1183.4759999765083, "tpot_ms": 7.3471465754029275, "ttft_ms": 1183.4763999795541, "visible_content_truncated": false, "visible_generation_ms": 3754.3915000278503 }, { "content_chunks": 45, "decode_tok_s": 168.38496771375787, "e2e_ms": 4937.972099985927, "effective_prefill_tok_s": 1910.4069576186398, "generation_ms": 3034.712699940428, "mean_inter_token_latency_ms": 5.938772406928431, "output_token_source": "usage", "output_tokens": 512, "planned_context_tokens": 4096, "prompt_tokens": 3636, "reasoning_chunks": 0, "request_canary": { "foreign_marker_count": 0, "marker": "ANVIL_REQ_0_00024", "passed": true }, "request_index": 24, "time_to_first_output_ms": 1903.2594000454992, "tpot_ms": 5.938772406928431, "ttft_ms": 1903.2599999336526, "visible_content_truncated": false, "visible_generation_ms": 3034.7121000522748 }, { "content_chunks": 45, "decode_tok_s": 151.29485300121388, "e2e_ms": 4938.411800074391, "effective_prefill_tok_s": 2330.0644947296964, "generation_ms": 3377.510800026357, "mean_inter_token_latency_ms": 6.609610176176824, "output_token_source": "usage", "output_tokens": 512, "planned_context_tokens": 4096, "prompt_tokens": 3637, "reasoning_chunks": 0, "request_canary": { "foreign_marker_count": 0, "marker": "ANVIL_REQ_0_00025", "passed": true }, "request_index": 25, "time_to_first_output_ms": 1560.901000048034, "tpot_ms": 6.609610176176824, "ttft_ms": 1560.9013000503182, "visible_content_truncated": false, "visible_generation_ms": 3377.510500024073 }, { "content_chunks": 45, "decode_tok_s": 122.45380849532096, "e2e_ms": 9758.101999992505, "effective_prefill_tok_s": 651.5550420116475, "generation_ms": 4173.002099967562, "mean_inter_token_latency_ms": 8.166344618331824, "output_token_source": "usage", "output_tokens": 512, "planned_context_tokens": 4096, "prompt_tokens": 3639, "reasoning_chunks": 0, "request_canary": { "foreign_marker_count": 0, "marker": "ANVIL_REQ_0_00026", "passed": true }, "request_index": 26, "time_to_first_output_ms": 5585.099900024943, "tpot_ms": 8.166344618331824, "ttft_ms": 5585.100500029512, "visible_content_truncated": false, "visible_generation_ms": 4173.001499962993 }, { "content_chunks": 45, "decode_tok_s": 263.35175352963586, "e2e_ms": 4937.794299912639, "effective_prefill_tok_s": 1214.042579323093, "generation_ms": 1940.3705999720842, "mean_inter_token_latency_ms": 3.7972027396713974, "output_token_source": "usage", "output_tokens": 512, "planned_context_tokens": 4096, "prompt_tokens": 3639, "reasoning_chunks": 0, "request_canary": { "foreign_marker_count": 0, "marker": "ANVIL_REQ_0_00027", "passed": true }, "request_index": 27, "time_to_first_output_ms": 2997.423699940555, "tpot_ms": 3.7972027396713974, "ttft_ms": 2997.424399945885, "visible_content_truncated": false, "visible_generation_ms": 1940.369899966754 }, { "content_chunks": 45, "decode_tok_s": 209.19077842686622, "e2e_ms": 4937.16490000952, "effective_prefill_tok_s": 1459.2578807789469, "generation_ms": 2442.746300017461, "mean_inter_token_latency_ms": 4.780325440347283, "output_token_source": "usage", "output_tokens": 512, "planned_context_tokens": 4096, "prompt_tokens": 3640, "reasoning_chunks": 0, "request_canary": { "foreign_marker_count": 0, "marker": "ANVIL_REQ_0_00028", "passed": true }, "request_index": 28, "time_to_first_output_ms": 2494.418599992059, "tpot_ms": 4.780325440347283, "ttft_ms": 2494.418999995105, "visible_content_truncated": false, "visible_generation_ms": 2442.7459000144154 }, { "content_chunks": 45, "decode_tok_s": 301.4668693613918, "e2e_ms": 4937.233499949798, "effective_prefill_tok_s": 1123.3154201208963, "generation_ms": 1695.0452999444678, "mean_inter_token_latency_ms": 3.317114089910896, "output_token_source": "usage", "output_tokens": 512, "planned_context_tokens": 4096, "prompt_tokens": 3642, "reasoning_chunks": 0, "request_canary": { "foreign_marker_count": 0, "marker": "ANVIL_REQ_0_00029", "passed": true }, "request_index": 29, "time_to_first_output_ms": 3242.18820000533, "tpot_ms": 3.317114089910896, "ttft_ms": 3242.188600008376, "visible_content_truncated": false, "visible_generation_ms": 1695.044899941422 }, { "content_chunks": 45, "decode_tok_s": 154.23965419186328, "e2e_ms": 4857.749400078319, "effective_prefill_tok_s": 2357.703803605601, "generation_ms": 3313.026100047864, "mean_inter_token_latency_ms": 6.4834170255339805, "output_token_source": "usage", "output_tokens": 512, "planned_context_tokens": 4096, "prompt_tokens": 3642, "reasoning_chunks": 0, "request_canary": { "foreign_marker_count": 0, "marker": "ANVIL_REQ_0_00030", "passed": true }, "request_index": 30, "time_to_first_output_ms": 1544.723300030455, "tpot_ms": 6.4834170255339805, "ttft_ms": 1544.7237000335008, "visible_content_truncated": false, "visible_generation_ms": 3313.025700044818 }, { "content_chunks": 45, "decode_tok_s": 139.17693051107125, "e2e_ms": 4857.686499948613, "effective_prefill_tok_s": 3071.407915667258, "generation_ms": 3671.5855000074953, "mean_inter_token_latency_ms": 7.185098825846371, "output_token_source": "usage", "output_tokens": 512, "planned_context_tokens": 4096, "prompt_tokens": 3643, "reasoning_chunks": 0, "request_canary": { "foreign_marker_count": 0, "marker": "ANVIL_REQ_0_00031", "passed": true }, "request_index": 31, "time_to_first_output_ms": 1186.100999941118, "tpot_ms": 7.185098825846371, "ttft_ms": 1186.1014999449253, "visible_content_truncated": false, "visible_generation_ms": 3671.585000003688 }, { "content_chunks": 45, "decode_tok_s": 180.05568618782587, "e2e_ms": 4856.691099936143, "effective_prefill_tok_s": 1805.1398136439711, "generation_ms": 2838.010899955407, "mean_inter_token_latency_ms": 5.553837377603536, "output_token_source": "usage", "output_tokens": 512, "planned_context_tokens": 4096, "prompt_tokens": 3644, "reasoning_chunks": 0, "request_canary": { "foreign_marker_count": 0, "marker": "ANVIL_REQ_0_00032", "passed": true }, "request_index": 32, "time_to_first_output_ms": 2018.6801999807358, "tpot_ms": 5.553837377603536, "ttft_ms": 2018.680899986066, "visible_content_truncated": false, "visible_generation_ms": 2838.0101999500766 }, { "content_chunks": 44, "decode_tok_s": 199.97731763763105, "e2e_ms": 4817.431000061333, "effective_prefill_tok_s": 1611.3052536133962, "generation_ms": 2555.289800046012, "mean_inter_token_latency_ms": 5.000567123377714, "output_token_source": "usage", "output_tokens": 512, "planned_context_tokens": 4096, "prompt_tokens": 3645, "reasoning_chunks": 0, "request_canary": { "foreign_marker_count": 0, "marker": "ANVIL_REQ_0_00033", "passed": true }, "request_index": 33, "time_to_first_output_ms": 2262.1412000153214, "tpot_ms": 5.000567123377714, "ttft_ms": 2262.1417000191286, "visible_content_truncated": false, "visible_generation_ms": 2555.2893000422046 }, { "content_chunks": 44, "decode_tok_s": 123.40011807086766, "e2e_ms": 9578.948900103569, "effective_prefill_tok_s": 670.4734947771415, "generation_ms": 4141.00090006832, "mean_inter_token_latency_ms": 8.103719960994756, "output_token_source": "usage", "output_tokens": 512, "planned_context_tokens": 4096, "prompt_tokens": 3646, "reasoning_chunks": 0, "request_canary": { "foreign_marker_count": 0, "marker": "ANVIL_REQ_0_00034", "passed": true }, "request_index": 34, "time_to_first_output_ms": 5437.948000035249, "tpot_ms": 8.103719960994756, "ttft_ms": 5437.948500039056, "visible_content_truncated": false, "visible_generation_ms": 4141.000400064513 }, { "content_chunks": 45, "decode_tok_s": 240.69878200461577, "e2e_ms": 4962.174099986441, "effective_prefill_tok_s": 1284.1696644015428, "generation_ms": 2122.985400026664, "mean_inter_token_latency_ms": 4.154570254455312, "output_token_source": "usage", "output_tokens": 512, "planned_context_tokens": 4096, "prompt_tokens": 3646, "reasoning_chunks": 0, "request_canary": { "foreign_marker_count": 0, "marker": "ANVIL_REQ_0_00035", "passed": true }, "request_index": 35, "time_to_first_output_ms": 2839.1886999597773, "tpot_ms": 4.154570254455312, "ttft_ms": 2839.1891999635845, "visible_content_truncated": false, "visible_generation_ms": 2122.984900022857 }, { "content_chunks": 45, "decode_tok_s": 285.0245215854015, "e2e_ms": 4961.719500017352, "effective_prefill_tok_s": 1150.8757021804909, "generation_ms": 1792.8282000357285, "mean_inter_token_latency_ms": 3.5084700587783337, "output_token_source": "usage", "output_tokens": 512, "planned_context_tokens": 4096, "prompt_tokens": 3647, "reasoning_chunks": 0, "request_canary": { "foreign_marker_count": 0, "marker": "ANVIL_REQ_0_00036", "passed": true }, "request_index": 36, "time_to_first_output_ms": 3168.891299981624, "tpot_ms": 3.5084700587783337, "ttft_ms": 3168.891799985431, "visible_content_truncated": false, "visible_generation_ms": 1792.8277000319213 }, { "content_chunks": 44, "decode_tok_s": 135.08849327231036, "e2e_ms": 4797.4195999559015, "effective_prefill_tok_s": 3596.087173573648, "generation_ms": 3782.705600024201, "mean_inter_token_latency_ms": 7.402554990262624, "output_token_source": "usage", "output_tokens": 512, "planned_context_tokens": 4096, "prompt_tokens": 3649, "reasoning_chunks": 0, "request_canary": { "foreign_marker_count": 0, "marker": "ANVIL_REQ_0_00037", "passed": true }, "request_index": 37, "time_to_first_output_ms": 1014.7139999317005, "tpot_ms": 7.402554990262624, "ttft_ms": 1014.7143999347463, "visible_content_truncated": false, "visible_generation_ms": 3782.705200021155 }, { "content_chunks": 44, "decode_tok_s": 268.1164332544987, "e2e_ms": 4797.059600008652, "effective_prefill_tok_s": 1262.1182723597553, "generation_ms": 1905.888400040567, "mean_inter_token_latency_ms": 3.7297228963611877, "output_token_source": "usage", "output_tokens": 512, "planned_context_tokens": 4096, "prompt_tokens": 3649, "reasoning_chunks": 0, "request_canary": { "foreign_marker_count": 0, "marker": "ANVIL_REQ_0_00038", "passed": true }, "request_index": 38, "time_to_first_output_ms": 2891.1711999680847, "tpot_ms": 3.7297228963611877, "ttft_ms": 2891.171499970369, "visible_content_truncated": false, "visible_generation_ms": 1905.8881000382826 }, { "content_chunks": 45, "decode_tok_s": 152.09133623293255, "e2e_ms": 4938.783000106923, "effective_prefill_tok_s": 2311.6484661850377, "generation_ms": 3359.8232000367716, "mean_inter_token_latency_ms": 6.574996477567067, "output_token_source": "usage", "output_tokens": 512, "planned_context_tokens": 4096, "prompt_tokens": 3650, "reasoning_chunks": 0, "request_canary": { "foreign_marker_count": 0, "marker": "ANVIL_REQ_0_00039", "passed": true }, "request_index": 39, "time_to_first_output_ms": 1578.9598000701517, "tpot_ms": 6.574996477567067, "ttft_ms": 1578.9604000747204, "visible_content_truncated": false, "visible_generation_ms": 3359.822600032203 }, { "content_chunks": 44, "decode_tok_s": 186.54611868994309, "e2e_ms": 4796.294300002046, "effective_prefill_tok_s": 1774.8930944205822, "generation_ms": 2739.2690000124276, "mean_inter_token_latency_ms": 5.36060469669751, "output_token_source": "usage", "output_tokens": 512, "planned_context_tokens": 4096, "prompt_tokens": 3651, "reasoning_chunks": 0, "request_canary": { "foreign_marker_count": 0, "marker": "ANVIL_REQ_0_00040", "passed": true }, "request_index": 40, "time_to_first_output_ms": 2057.0252999896184, "tpot_ms": 5.36060469669751, "ttft_ms": 2057.0257999934256, "visible_content_truncated": false, "visible_generation_ms": 2739.2685000086203 }, { "content_chunks": 45, "decode_tok_s": 202.38956728752441, "e2e_ms": 4938.591299927793, "effective_prefill_tok_s": 1512.9936825429127, "generation_ms": 2524.833699921146, "mean_inter_token_latency_ms": 4.940966144659777, "output_token_source": "usage", "output_tokens": 512, "planned_context_tokens": 4096, "prompt_tokens": 3652, "reasoning_chunks": 0, "request_canary": { "foreign_marker_count": 0, "marker": "ANVIL_REQ_0_00041", "passed": true }, "request_index": 41, "time_to_first_output_ms": 2413.7576000066474, "tpot_ms": 4.940966144659777, "ttft_ms": 2413.758200011216, "visible_content_truncated": false, "visible_generation_ms": 2524.8330999165773 }, { "content_chunks": 44, "decode_tok_s": 123.66661194873251, "e2e_ms": 9405.248600058258, "effective_prefill_tok_s": 692.7520067388192, "generation_ms": 4132.07729998976, "mean_inter_token_latency_ms": 8.086256947142388, "output_token_source": "usage", "output_tokens": 512, "planned_context_tokens": 4096, "prompt_tokens": 3653, "reasoning_chunks": 0, "request_canary": { "foreign_marker_count": 0, "marker": "ANVIL_REQ_0_00042", "passed": true }, "request_index": 42, "time_to_first_output_ms": 5273.171300068498, "tpot_ms": 8.086256947142388, "ttft_ms": 5273.171900073066, "visible_content_truncated": false, "visible_generation_ms": 4132.076699985191 }, { "content_chunks": 45, "decode_tok_s": 283.7025436064164, "e2e_ms": 4797.04740003217, "effective_prefill_tok_s": 1219.3472930148669, "generation_ms": 1801.1822999687865, "mean_inter_token_latency_ms": 3.52481859093696, "output_token_source": "usage", "output_tokens": 512, "planned_context_tokens": 4096, "prompt_tokens": 3653, "reasoning_chunks": 0, "request_canary": { "foreign_marker_count": 0, "marker": "ANVIL_REQ_0_00043", "passed": true }, "request_index": 43, "time_to_first_output_ms": 2995.8651000633836, "tpot_ms": 3.52481859093696, "ttft_ms": 2995.865600067191, "visible_content_truncated": false, "visible_generation_ms": 1801.1817999649793 }, { "content_chunks": 45, "decode_tok_s": 147.8079787460871, "e2e_ms": 4828.729700064287, "effective_prefill_tok_s": 2664.155623483677, "generation_ms": 3457.188200089149, "mean_inter_token_latency_ms": 6.765534638139235, "output_token_source": "usage", "output_tokens": 512, "planned_context_tokens": 4096, "prompt_tokens": 3654, "reasoning_chunks": 0, "request_canary": { "foreign_marker_count": 0, "marker": "ANVIL_REQ_0_00044", "passed": true }, "request_index": 44, "time_to_first_output_ms": 1371.5414999751374, "tpot_ms": 6.765534638139235, "ttft_ms": 1371.5418999781832, "visible_content_truncated": false, "visible_generation_ms": 3457.1878000861034 }, { "content_chunks": 45, "decode_tok_s": 138.39660143394252, "e2e_ms": 4827.858699951321, "effective_prefill_tok_s": 3216.8824245782635, "generation_ms": 3692.287200014107, "mean_inter_token_latency_ms": 7.225610958931716, "output_token_source": "usage", "output_tokens": 512, "planned_context_tokens": 4096, "prompt_tokens": 3653, "reasoning_chunks": 0, "request_canary": { "foreign_marker_count": 0, "marker": "ANVIL_REQ_0_00045", "passed": true }, "request_index": 45, "time_to_first_output_ms": 1135.571499937214, "tpot_ms": 7.225610958931716, "ttft_ms": 1135.572199942544, "visible_content_truncated": false, "visible_generation_ms": 3692.286500008777 }, { "content_chunks": 45, "decode_tok_s": 203.96136070084793, "e2e_ms": 4933.919799979776, "effective_prefill_tok_s": 1504.605662156382, "generation_ms": 2505.3764999611303, "mean_inter_token_latency_ms": 4.902889432409257, "output_token_source": "usage", "output_tokens": 512, "planned_context_tokens": 4096, "prompt_tokens": 3654, "reasoning_chunks": 0, "request_canary": { "foreign_marker_count": 0, "marker": "ANVIL_REQ_0_00046", "passed": true }, "request_index": 46, "time_to_first_output_ms": 2428.543300018646, "tpot_ms": 4.902889432409257, "ttft_ms": 2428.5437000216916, "visible_content_truncated": false, "visible_generation_ms": 2505.3760999580845 }, { "content_chunks": 45, "decode_tok_s": 171.22694392385722, "e2e_ms": 4934.192199958488, "effective_prefill_tok_s": 1874.5043205725574, "generation_ms": 2984.3433999922127, "mean_inter_token_latency_ms": 5.840202348321355, "output_token_source": "usage", "output_tokens": 512, "planned_context_tokens": 4096, "prompt_tokens": 3655, "reasoning_chunks": 0, "request_canary": { "foreign_marker_count": 0, "marker": "ANVIL_REQ_0_00047", "passed": true }, "request_index": 47, "time_to_first_output_ms": 1949.8487999662757, "tpot_ms": 5.840202348321355, "ttft_ms": 1949.849299970083, "visible_content_truncated": false, "visible_generation_ms": 2984.3428999884054 }, { "content_chunks": 45, "decode_tok_s": 284.14501406071025, "e2e_ms": 4791.852599941194, "effective_prefill_tok_s": 1220.9889435755863, "generation_ms": 1798.3774999156594, "mean_inter_token_latency_ms": 3.519329745431819, "output_token_source": "usage", "output_tokens": 512, "planned_context_tokens": 4096, "prompt_tokens": 3655, "reasoning_chunks": 0, "request_canary": { "foreign_marker_count": 0, "marker": "ANVIL_REQ_0_00048", "passed": true }, "request_index": 48, "time_to_first_output_ms": 2993.4751000255346, "tpot_ms": 3.519329745431819, "ttft_ms": 2993.475600029342, "visible_content_truncated": false, "visible_generation_ms": 1798.3769999118522 }, { "content_chunks": 44, "decode_tok_s": 123.11921247995195, "e2e_ms": 9421.03229998611, "effective_prefill_tok_s": 693.4716183362633, "generation_ms": 4150.4488999489695, "mean_inter_token_latency_ms": 8.122209197551799, "output_token_source": "usage", "output_tokens": 512, "planned_context_tokens": 4096, "prompt_tokens": 3655, "reasoning_chunks": 0, "request_canary": { "foreign_marker_count": 0, "marker": "ANVIL_REQ_0_00049", "passed": true }, "request_index": 49, "time_to_first_output_ms": 5270.58340003714, "tpot_ms": 8.122209197551799, "ttft_ms": 5270.583800040185, "visible_content_truncated": false, "visible_generation_ms": 4150.448499945924 }, { "content_chunks": 44, "decode_tok_s": 255.31331490930023, "e2e_ms": 4647.491200012155, "effective_prefill_tok_s": 1381.315275150201, "generation_ms": 2001.4623999595642, "mean_inter_token_latency_ms": 3.916756164304431, "output_token_source": "usage", "output_tokens": 512, "planned_context_tokens": 4096, "prompt_tokens": 3655, "reasoning_chunks": 0, "request_canary": { "foreign_marker_count": 0, "marker": "ANVIL_REQ_0_00050", "passed": true }, "request_index": 50, "time_to_first_output_ms": 2646.0288000525907, "tpot_ms": 3.916756164304431, "ttft_ms": 2646.0292000556365, "visible_content_truncated": false, "visible_generation_ms": 2001.4619999565184 }, { "content_chunks": 45, "decode_tok_s": 132.51529300137676, "e2e_ms": 4852.758500026539, "effective_prefill_tok_s": 3667.4701321141147, "generation_ms": 3856.158700073138, "mean_inter_token_latency_ms": 7.546298825974829, "output_token_source": "usage", "output_tokens": 512, "planned_context_tokens": 4096, "prompt_tokens": 3655, "reasoning_chunks": 0, "request_canary": { "foreign_marker_count": 0, "marker": "ANVIL_REQ_0_00051", "passed": true }, "request_index": 51, "time_to_first_output_ms": 996.5997999534011, "tpot_ms": 7.546298825974829, "ttft_ms": 996.6001999564469, "visible_content_truncated": false, "visible_generation_ms": 3856.158300070092 }, { "content_chunks": 45, "decode_tok_s": 150.71139761560457, "e2e_ms": 4957.499300013296, "effective_prefill_tok_s": 2332.611957284618, "generation_ms": 3390.5862999381498, "mean_inter_token_latency_ms": 6.635198238626516, "output_token_source": "usage", "output_tokens": 512, "planned_context_tokens": 4096, "prompt_tokens": 3655, "reasoning_chunks": 0, "request_canary": { "foreign_marker_count": 0, "marker": "ANVIL_REQ_0_00052", "passed": true }, "request_index": 52, "time_to_first_output_ms": 1566.9130000751466, "tpot_ms": 6.635198238626516, "ttft_ms": 1566.9134000781924, "visible_content_truncated": false, "visible_generation_ms": 3390.585899935104 }, { "content_chunks": 44, "decode_tok_s": 212.89677856066902, "e2e_ms": 4813.410200062208, "effective_prefill_tok_s": 1515.0096179415311, "generation_ms": 2400.2241999842227, "mean_inter_token_latency_ms": 4.697111937346816, "output_token_source": "usage", "output_tokens": 512, "planned_context_tokens": 4096, "prompt_tokens": 3656, "reasoning_chunks": 0, "request_canary": { "foreign_marker_count": 0, "marker": "ANVIL_REQ_0_00053", "passed": true }, "request_index": 53, "time_to_first_output_ms": 2413.186000077985, "tpot_ms": 4.697111937346816, "ttft_ms": 2413.186600082554, "visible_content_truncated": false, "visible_generation_ms": 2400.223599979654 }, { "content_chunks": 45, "decode_tok_s": 175.71818186789196, "e2e_ms": 4957.131500006653, "effective_prefill_tok_s": 1784.7156600209596, "generation_ms": 2908.065600087866, "mean_inter_token_latency_ms": 5.690930724242399, "output_token_source": "usage", "output_tokens": 512, "planned_context_tokens": 4096, "prompt_tokens": 3657, "reasoning_chunks": 0, "request_canary": { "foreign_marker_count": 0, "marker": "ANVIL_REQ_0_00054", "passed": true }, "request_index": 54, "time_to_first_output_ms": 2049.065899918787, "tpot_ms": 5.690930724242399, "ttft_ms": 2049.0663999225944, "visible_content_truncated": false, "visible_generation_ms": 2908.065100084059 }, { "content_chunks": 45, "decode_tok_s": 247.67460926398547, "e2e_ms": 4814.038399956189, "effective_prefill_tok_s": 1329.772006658941, "generation_ms": 2063.190900022164, "mean_inter_token_latency_ms": 4.037555577342786, "output_token_source": "usage", "output_tokens": 512, "planned_context_tokens": 4096, "prompt_tokens": 3658, "reasoning_chunks": 0, "request_canary": { "foreign_marker_count": 0, "marker": "ANVIL_REQ_0_00055", "passed": true }, "request_index": 55, "time_to_first_output_ms": 2750.847499934025, "tpot_ms": 4.037555577342786, "ttft_ms": 2750.847899937071, "visible_content_truncated": false, "visible_generation_ms": 2063.190500019118 }, { "content_chunks": 44, "decode_tok_s": 305.9929528291385, "e2e_ms": 4669.055299949832, "effective_prefill_tok_s": 1220.039917546931, "generation_ms": 1669.973099953495, "mean_inter_token_latency_ms": 3.268049119282769, "output_token_source": "usage", "output_tokens": 512, "planned_context_tokens": 4096, "prompt_tokens": 3659, "reasoning_chunks": 0, "request_canary": { "foreign_marker_count": 0, "marker": "ANVIL_REQ_0_00056", "passed": true }, "request_index": 56, "time_to_first_output_ms": 2999.0821999963373, "tpot_ms": 3.268049119282769, "ttft_ms": 2999.0824999986216, "visible_content_truncated": false, "visible_generation_ms": 1669.9727999512106 }, { "content_chunks": 44, "decode_tok_s": 122.73656127153755, "e2e_ms": 9458.970800042152, "effective_prefill_tok_s": 691.1421373016464, "generation_ms": 4163.388599990867, "mean_inter_token_latency_ms": 8.147531506831442, "output_token_source": "usage", "output_tokens": 512, "planned_context_tokens": 4096, "prompt_tokens": 3660, "reasoning_chunks": 0, "request_canary": { "foreign_marker_count": 0, "marker": "ANVIL_REQ_0_00057", "passed": true }, "request_index": 57, "time_to_first_output_ms": 5295.582200051285, "tpot_ms": 8.147531506831442, "ttft_ms": 5295.582700055093, "visible_content_truncated": false, "visible_generation_ms": 4163.38809998706 }, { "content_chunks": 44, "decode_tok_s": 133.43554861879582, "e2e_ms": 4829.321599914692, "effective_prefill_tok_s": 3661.888374489671, "generation_ms": 3829.5642000157386, "mean_inter_token_latency_ms": 7.494254794551347, "output_token_source": "usage", "output_tokens": 512, "planned_context_tokens": 4096, "prompt_tokens": 3661, "reasoning_chunks": 0, "request_canary": { "foreign_marker_count": 0, "marker": "ANVIL_REQ_0_00058", "passed": true }, "request_index": 58, "time_to_first_output_ms": 999.7573998989537, "tpot_ms": 7.494254794551347, "ttft_ms": 999.7577999019995, "visible_content_truncated": false, "visible_generation_ms": 3829.563800012693 }, { "content_chunks": 44, "decode_tok_s": 211.821189256492, "e2e_ms": 4828.427700093016, "effective_prefill_tok_s": 1515.3047852290294, "generation_ms": 2412.412100005895, "mean_inter_token_latency_ms": 4.720963013710167, "output_token_source": "usage", "output_tokens": 512, "planned_context_tokens": 4096, "prompt_tokens": 3661, "reasoning_chunks": 0, "request_canary": { "foreign_marker_count": 0, "marker": "ANVIL_REQ_0_00059", "passed": true }, "request_index": 59, "time_to_first_output_ms": 2416.015600087121, "tpot_ms": 4.720963013710167, "ttft_ms": 2416.0161000909284, "visible_content_truncated": false, "visible_generation_ms": 2412.4116000020877 }, { "content_chunks": 44, "decode_tok_s": 184.15893658922545, "e2e_ms": 4828.054800047539, "effective_prefill_tok_s": 1783.489880514668, "generation_ms": 2774.777099955827, "mean_inter_token_latency_ms": 5.4300921721249065, "output_token_source": "usage", "output_tokens": 512, "planned_context_tokens": 4096, "prompt_tokens": 3662, "reasoning_chunks": 0, "request_canary": { "foreign_marker_count": 0, "marker": "ANVIL_REQ_0_00060", "passed": true }, "request_index": 60, "time_to_first_output_ms": 2053.277700091712, "tpot_ms": 5.4300921721249065, "ttft_ms": 2053.2782000955194, "visible_content_truncated": false, "visible_generation_ms": 2774.7765999520198 }, { "content_chunks": 44, "decode_tok_s": 156.81935244941275, "e2e_ms": 4827.673000050709, "effective_prefill_tok_s": 2334.3899161942145, "generation_ms": 3258.526400080882, "mean_inter_token_latency_ms": 6.376763992330493, "output_token_source": "usage", "output_tokens": 512, "planned_context_tokens": 4096, "prompt_tokens": 3663, "reasoning_chunks": 0, "request_canary": { "foreign_marker_count": 0, "marker": "ANVIL_REQ_0_00061", "passed": true }, "request_index": 61, "time_to_first_output_ms": 1569.1465999698266, "tpot_ms": 6.376763992330493, "ttft_ms": 1569.1469999728724, "visible_content_truncated": false, "visible_generation_ms": 3258.526000077836 }, { "content_chunks": 45, "decode_tok_s": 246.46035185971866, "e2e_ms": 4826.131999958307, "effective_prefill_tok_s": 1331.3832050946296, "generation_ms": 2073.3558000065386, "mean_inter_token_latency_ms": 4.057447749523559, "output_token_source": "usage", "output_tokens": 512, "planned_context_tokens": 4096, "prompt_tokens": 3665, "reasoning_chunks": 0, "request_canary": { "foreign_marker_count": 0, "marker": "ANVIL_REQ_0_00062", "passed": true }, "request_index": 62, "time_to_first_output_ms": 2752.776199951768, "tpot_ms": 4.057447749523559, "ttft_ms": 2752.7765999548137, "visible_content_truncated": false, "visible_generation_ms": 2073.355400003493 }, { "content_chunks": 44, "decode_tok_s": 122.38419060669175, "e2e_ms": 9484.733500052243, "effective_prefill_tok_s": 690.2906671727822, "generation_ms": 4175.375899998471, "mean_inter_token_latency_ms": 8.17099001956648, "output_token_source": "usage", "output_tokens": 512, "planned_context_tokens": 4096, "prompt_tokens": 3665, "reasoning_chunks": 0, "request_canary": { "foreign_marker_count": 0, "marker": "ANVIL_REQ_0_00063", "passed": true }, "request_index": 63, "time_to_first_output_ms": 5309.357600053772, "tpot_ms": 8.17099001956648, "ttft_ms": 5309.35810005758, "visible_content_truncated": false, "visible_generation_ms": 4175.375399994664 }, { "content_chunks": 45, "decode_tok_s": 281.5250792025509, "e2e_ms": 4825.420800014399, "effective_prefill_tok_s": 1217.815909278437, "generation_ms": 1815.1135999942198, "mean_inter_token_latency_ms": 3.5520814089906456, "output_token_source": "usage", "output_tokens": 512, "planned_context_tokens": 4096, "prompt_tokens": 3666, "reasoning_chunks": 0, "request_canary": { "foreign_marker_count": 0, "marker": "ANVIL_REQ_0_00064", "passed": true }, "request_index": 64, "time_to_first_output_ms": 3010.307200020179, "tpot_ms": 3.5520814089906456, "ttft_ms": 3010.307600023225, "visible_content_truncated": false, "visible_generation_ms": 1815.113199991174 }, { "content_chunks": 45, "decode_tok_s": 136.9586364827062, "e2e_ms": 4879.747899947688, "effective_prefill_tok_s": 3192.3199069074376, "generation_ms": 3731.053499970585, "mean_inter_token_latency_ms": 7.301474559629325, "output_token_source": "usage", "output_tokens": 512, "planned_context_tokens": 4096, "prompt_tokens": 3667, "reasoning_chunks": 0, "request_canary": { "foreign_marker_count": 0, "marker": "ANVIL_REQ_0_00065", "passed": true }, "request_index": 65, "time_to_first_output_ms": 1148.6943999771029, "tpot_ms": 7.301474559629325, "ttft_ms": 1148.6948999809101, "visible_content_truncated": false, "visible_generation_ms": 3731.052999966778 }, { "content_chunks": 44, "decode_tok_s": 148.17390164979736, "e2e_ms": 4838.861499913037, "effective_prefill_tok_s": 2637.0097776698826, "generation_ms": 3448.6504999222234, "mean_inter_token_latency_ms": 6.74882681002392, "output_token_source": "usage", "output_tokens": 512, "planned_context_tokens": 4096, "prompt_tokens": 3666, "reasoning_chunks": 0, "request_canary": { "foreign_marker_count": 0, "marker": "ANVIL_REQ_0_00066", "passed": true }, "request_index": 66, "time_to_first_output_ms": 1390.2109999908134, "tpot_ms": 6.74882681002392, "ttft_ms": 1390.2114999946207, "visible_content_truncated": false, "visible_generation_ms": 3448.649999918416 }, { "content_chunks": 44, "decode_tok_s": 253.66963407320102, "e2e_ms": 4841.807899996638, "effective_prefill_tok_s": 1296.9619047559133, "generation_ms": 2014.4310999894515, "mean_inter_token_latency_ms": 3.942135225028281, "output_token_source": "usage", "output_tokens": 512, "planned_context_tokens": 4096, "prompt_tokens": 3667, "reasoning_chunks": 0, "request_canary": { "foreign_marker_count": 0, "marker": "ANVIL_REQ_0_00067", "passed": true }, "request_index": 67, "time_to_first_output_ms": 2827.376800007187, "tpot_ms": 3.942135225028281, "ttft_ms": 2827.377300010994, "visible_content_truncated": false, "visible_generation_ms": 2014.4305999856442 }, { "content_chunks": 44, "decode_tok_s": 178.3477335509409, "e2e_ms": 4841.446099919267, "effective_prefill_tok_s": 1856.0340004366615, "generation_ms": 2865.1891999179497, "mean_inter_token_latency_ms": 5.607023874594813, "output_token_source": "usage", "output_tokens": 512, "planned_context_tokens": 4096, "prompt_tokens": 3668, "reasoning_chunks": 0, "request_canary": { "foreign_marker_count": 0, "marker": "ANVIL_REQ_0_00068", "passed": true }, "request_index": 68, "time_to_first_output_ms": 1976.2569000013173, "tpot_ms": 5.607023874594813, "ttft_ms": 1976.257500005886, "visible_content_truncated": false, "visible_generation_ms": 2865.188599913381 }, { "content_chunks": 44, "decode_tok_s": 214.8144539065477, "e2e_ms": 4840.552799985744, "effective_prefill_tok_s": 1490.3997176014798, "generation_ms": 2378.797100041993, "mean_inter_token_latency_ms": 4.6551802349158375, "output_token_source": "usage", "output_tokens": 512, "planned_context_tokens": 4096, "prompt_tokens": 3669, "reasoning_chunks": 0, "request_canary": { "foreign_marker_count": 0, "marker": "ANVIL_REQ_0_00069", "passed": true }, "request_index": 69, "time_to_first_output_ms": 2461.755699943751, "tpot_ms": 4.6551802349158375, "ttft_ms": 2461.756099946797, "visible_content_truncated": false, "visible_generation_ms": 2378.796700038947 }, { "content_chunks": 45, "decode_tok_s": 120.93051940597421, "e2e_ms": 9549.797599902377, "effective_prefill_tok_s": 689.1136404096146, "generation_ms": 4225.566899986006, "mean_inter_token_latency_ms": 8.26921115457144, "output_token_source": "usage", "output_tokens": 512, "planned_context_tokens": 4096, "prompt_tokens": 3669, "reasoning_chunks": 0, "request_canary": { "foreign_marker_count": 0, "marker": "ANVIL_REQ_0_00070", "passed": true }, "request_index": 70, "time_to_first_output_ms": 5324.23069991637, "tpot_ms": 8.26921115457144, "ttft_ms": 5324.231299920939, "visible_content_truncated": false, "visible_generation_ms": 4225.566299981438 }, { "content_chunks": 45, "decode_tok_s": 281.8593450134762, "e2e_ms": 4839.2848999938, "effective_prefill_tok_s": 1212.6924021679013, "generation_ms": 1812.9610000178218, "mean_inter_token_latency_ms": 3.5478688845749935, "output_token_source": "usage", "output_tokens": 512, "planned_context_tokens": 4096, "prompt_tokens": 3670, "reasoning_chunks": 0, "request_canary": { "foreign_marker_count": 0, "marker": "ANVIL_REQ_0_00071", "passed": true }, "request_index": 71, "time_to_first_output_ms": 3026.323899975978, "tpot_ms": 3.5478688845749935, "ttft_ms": 3026.324199978262, "visible_content_truncated": false, "visible_generation_ms": 1812.9607000155374 }, { "content_chunks": 45, "decode_tok_s": 146.29224174321834, "e2e_ms": 4891.652099904604, "effective_prefill_tok_s": 2624.6852398617775, "generation_ms": 3493.0081999627873, "mean_inter_token_latency_ms": 6.835632485250073, "output_token_source": "usage", "output_tokens": 512, "planned_context_tokens": 4096, "prompt_tokens": 3671, "reasoning_chunks": 0, "request_canary": { "foreign_marker_count": 0, "marker": "ANVIL_REQ_0_00072", "passed": true }, "request_index": 72, "time_to_first_output_ms": 1398.643899941817, "tpot_ms": 6.835632485250073, "ttft_ms": 1398.6444999463856, "visible_content_truncated": false, "visible_generation_ms": 3493.0075999582186 }, { "content_chunks": 45, "decode_tok_s": 136.53478457711682, "e2e_ms": 4891.93649997469, "effective_prefill_tok_s": 3194.986863604168, "generation_ms": 3742.6359999226406, "mean_inter_token_latency_ms": 7.3241409000443065, "output_token_source": "usage", "output_tokens": 512, "planned_context_tokens": 4096, "prompt_tokens": 3672, "reasoning_chunks": 0, "request_canary": { "foreign_marker_count": 0, "marker": "ANVIL_REQ_0_00073", "passed": true }, "request_index": 73, "time_to_first_output_ms": 1149.3005000520498, "tpot_ms": 7.3241409000443065, "ttft_ms": 1149.3009000550956, "visible_content_truncated": false, "visible_generation_ms": 3742.635599919595 }, { "content_chunks": 44, "decode_tok_s": 214.41665411678142, "e2e_ms": 4851.366499904543, "effective_prefill_tok_s": 1487.75030881284, "generation_ms": 2383.210399886593, "mean_inter_token_latency_ms": 4.6638168295236655, "output_token_source": "usage", "output_tokens": 512, "planned_context_tokens": 4096, "prompt_tokens": 3672, "reasoning_chunks": 0, "request_canary": { "foreign_marker_count": 0, "marker": "ANVIL_REQ_0_00074", "passed": true }, "request_index": 74, "time_to_first_output_ms": 2468.15610001795, "tpot_ms": 4.6638168295236655, "ttft_ms": 2468.156599905342, "visible_content_truncated": false, "visible_generation_ms": 2383.2098999992013 }, { "content_chunks": 44, "decode_tok_s": 304.5474960587057, "e2e_ms": 4850.784300011583, "effective_prefill_tok_s": 1157.3063266585843, "generation_ms": 1677.8992000035942, "mean_inter_token_latency_ms": 3.28356007828492, "output_token_source": "usage", "output_tokens": 512, "planned_context_tokens": 4096, "prompt_tokens": 3672, "reasoning_chunks": 0, "request_canary": { "foreign_marker_count": 0, "marker": "ANVIL_REQ_0_00075", "passed": true }, "request_index": 75, "time_to_first_output_ms": 3172.8851000079885, "tpot_ms": 3.28356007828492, "ttft_ms": 3172.8858000133187, "visible_content_truncated": false, "visible_generation_ms": 1677.898499998264 }, { "content_chunks": 44, "decode_tok_s": 178.22301055452428, "e2e_ms": 4849.789900006726, "effective_prefill_tok_s": 1852.1174969234482, "generation_ms": 2867.1943000517786, "mean_inter_token_latency_ms": 5.610947749612091, "output_token_source": "usage", "output_tokens": 512, "planned_context_tokens": 4096, "prompt_tokens": 3672, "reasoning_chunks": 0, "request_canary": { "foreign_marker_count": 0, "marker": "ANVIL_REQ_0_00076", "passed": true }, "request_index": 76, "time_to_first_output_ms": 1982.5955999549478, "tpot_ms": 5.610947749612091, "ttft_ms": 1982.5959999579936, "visible_content_truncated": false, "visible_generation_ms": 2867.1939000487328 }, { "content_chunks": 44, "decode_tok_s": 252.9442262468486, "e2e_ms": 4850.913899950683, "effective_prefill_tok_s": 1297.2030260971462, "generation_ms": 2020.2081999741495, "mean_inter_token_latency_ms": 3.9534407044503905, "output_token_source": "usage", "output_tokens": 512, "planned_context_tokens": 4096, "prompt_tokens": 3672, "reasoning_chunks": 0, "request_canary": { "foreign_marker_count": 0, "marker": "ANVIL_REQ_0_00077", "passed": true }, "request_index": 77, "time_to_first_output_ms": 2830.7056999765337, "tpot_ms": 3.9534407044503905, "ttft_ms": 2830.7062999811023, "visible_content_truncated": false, "visible_generation_ms": 2020.2075999695808 }, { "content_chunks": 45, "decode_tok_s": 120.67526192718665, "e2e_ms": 9573.564899968915, "effective_prefill_tok_s": 687.7615289503663, "generation_ms": 4234.504999942146, "mean_inter_token_latency_ms": 8.286702543918095, "output_token_source": "usage", "output_tokens": 512, "planned_context_tokens": 4096, "prompt_tokens": 3672, "reasoning_chunks": 0, "request_canary": { "foreign_marker_count": 0, "marker": "ANVIL_REQ_0_00078", "passed": true }, "request_index": 78, "time_to_first_output_ms": 5339.059900026768, "tpot_ms": 8.286702543918095, "ttft_ms": 5339.060400030576, "visible_content_truncated": false, "visible_generation_ms": 4234.504499938339 }, { "content_chunks": 44, "decode_tok_s": 152.75911053791032, "e2e_ms": 4824.656300013885, "effective_prefill_tok_s": 2483.2373034388334, "generation_ms": 3345.136000076309, "mean_inter_token_latency_ms": 6.5462544032804475, "output_token_source": "usage", "output_tokens": 512, "planned_context_tokens": 4096, "prompt_tokens": 3674, "reasoning_chunks": 0, "request_canary": { "foreign_marker_count": 0, "marker": "ANVIL_REQ_0_00079", "passed": true }, "request_index": 79, "time_to_first_output_ms": 1479.520299937576, "tpot_ms": 6.5462544032804475, "ttft_ms": 1479.5205999398604, "visible_content_truncated": false, "visible_generation_ms": 3345.1357000740245 }, { "content_chunks": 44, "decode_tok_s": 254.76081151188413, "e2e_ms": 4823.783900006674, "effective_prefill_tok_s": 1303.7703697441364, "generation_ms": 2005.8029999490827, "mean_inter_token_latency_ms": 3.9252504891371482, "output_token_source": "usage", "output_tokens": 512, "planned_context_tokens": 4096, "prompt_tokens": 3674, "reasoning_chunks": 0, "request_canary": { "foreign_marker_count": 0, "marker": "ANVIL_REQ_0_00080", "passed": true }, "request_index": 80, "time_to_first_output_ms": 2817.9809000575915, "tpot_ms": 3.9252504891371482, "ttft_ms": 2817.9814000613987, "visible_content_truncated": false, "visible_generation_ms": 2005.8024999452755 }, { "content_chunks": 45, "decode_tok_s": 176.29289709253268, "e2e_ms": 4863.456999999471, "effective_prefill_tok_s": 1870.351127724293, "generation_ms": 2898.585299961269, "mean_inter_token_latency_ms": 5.672378277810703, "output_token_source": "usage", "output_tokens": 512, "planned_context_tokens": 4096, "prompt_tokens": 3675, "reasoning_chunks": 0, "request_canary": { "foreign_marker_count": 0, "marker": "ANVIL_REQ_0_00081", "passed": true }, "request_index": 81, "time_to_first_output_ms": 1964.871700038202, "tpot_ms": 5.672378277810703, "ttft_ms": 1964.8723000427708, "visible_content_truncated": false, "visible_generation_ms": 2898.5846999567 }, { "content_chunks": 44, "decode_tok_s": 205.03522372354706, "e2e_ms": 4823.515600059181, "effective_prefill_tok_s": 1576.8290884899527, "generation_ms": 2492.2547000460327, "mean_inter_token_latency_ms": 4.877210763299477, "output_token_source": "usage", "output_tokens": 512, "planned_context_tokens": 4096, "prompt_tokens": 3676, "reasoning_chunks": 0, "request_canary": { "foreign_marker_count": 0, "marker": "ANVIL_REQ_0_00082", "passed": true }, "request_index": 82, "time_to_first_output_ms": 2331.260900013149, "tpot_ms": 4.877210763299477, "ttft_ms": 2331.2613000161946, "visible_content_truncated": false, "visible_generation_ms": 2492.254300042987 }, { "content_chunks": 45, "decode_tok_s": 136.36872138362594, "e2e_ms": 4862.1823999565095, "effective_prefill_tok_s": 3297.7909732618327, "generation_ms": 3747.1935999346897, "mean_inter_token_latency_ms": 7.333059882455362, "output_token_source": "usage", "output_tokens": 512, "planned_context_tokens": 4096, "prompt_tokens": 3677, "reasoning_chunks": 0, "request_canary": { "foreign_marker_count": 0, "marker": "ANVIL_REQ_0_00083", "passed": true }, "request_index": 83, "time_to_first_output_ms": 1114.9888000218198, "tpot_ms": 7.333059882455362, "ttft_ms": 1114.9892000248656, "visible_content_truncated": false, "visible_generation_ms": 3747.193199931644 }, { "content_chunks": 45, "decode_tok_s": 298.35353222805566, "e2e_ms": 4862.661699997261, "effective_prefill_tok_s": 1167.3280838167525, "generation_ms": 1712.733200052753, "mean_inter_token_latency_ms": 3.3517283758370904, "output_token_source": "usage", "output_tokens": 512, "planned_context_tokens": 4096, "prompt_tokens": 3677, "reasoning_chunks": 0, "request_canary": { "foreign_marker_count": 0, "marker": "ANVIL_REQ_0_00084", "passed": true }, "request_index": 84, "time_to_first_output_ms": 3149.928499944508, "tpot_ms": 3.3517283758370904, "ttft_ms": 3149.9290999490768, "visible_content_truncated": false, "visible_generation_ms": 1712.7326000481844 }, { "content_chunks": 44, "decode_tok_s": 121.27403044316081, "e2e_ms": 9703.999000019394, "effective_prefill_tok_s": 669.8964124839441, "generation_ms": 4213.597900001332, "mean_inter_token_latency_ms": 8.245788454014349, "output_token_source": "usage", "output_tokens": 512, "planned_context_tokens": 4096, "prompt_tokens": 3678, "reasoning_chunks": 0, "request_canary": { "foreign_marker_count": 0, "marker": "ANVIL_REQ_0_00085", "passed": true }, "request_index": 85, "time_to_first_output_ms": 5490.401100018062, "tpot_ms": 8.245788454014349, "ttft_ms": 5490.4015000211075, "visible_content_truncated": false, "visible_generation_ms": 4213.597499998286 }, { "content_chunks": 44, "decode_tok_s": 152.09602610600174, "e2e_ms": 4881.562800030224, "effective_prefill_tok_s": 2417.463244622569, "generation_ms": 3359.7196000628173, "mean_inter_token_latency_ms": 6.57479373789201, "output_token_source": "usage", "output_tokens": 512, "planned_context_tokens": 4096, "prompt_tokens": 3679, "reasoning_chunks": 0, "request_canary": { "foreign_marker_count": 0, "marker": "ANVIL_REQ_0_00086", "passed": true }, "request_index": 86, "time_to_first_output_ms": 1521.8431999674067, "tpot_ms": 6.57479373789201, "ttft_ms": 1521.843699971214, "visible_content_truncated": false, "visible_generation_ms": 3359.71910005901 }, { "content_chunks": 45, "decode_tok_s": 167.2791725079638, "e2e_ms": 4919.881799956784, "effective_prefill_tok_s": 1973.0758783863882, "generation_ms": 3054.773599957116, "mean_inter_token_latency_ms": 5.9780305282918125, "output_token_source": "usage", "output_tokens": 512, "planned_context_tokens": 4096, "prompt_tokens": 3680, "reasoning_chunks": 0, "request_canary": { "foreign_marker_count": 0, "marker": "ANVIL_REQ_0_00087", "passed": true }, "request_index": 87, "time_to_first_output_ms": 1865.1081999996677, "tpot_ms": 5.9780305282918125, "ttft_ms": 1865.108500001952, "visible_content_truncated": false, "visible_generation_ms": 3054.7732999548316 }, { "content_chunks": 45, "decode_tok_s": 135.76699344561442, "e2e_ms": 4919.215100002475, "effective_prefill_tok_s": 3184.1408838142206, "generation_ms": 3763.801399967633, "mean_inter_token_latency_ms": 7.365560469603979, "output_token_source": "usage", "output_tokens": 512, "planned_context_tokens": 4096, "prompt_tokens": 3679, "reasoning_chunks": 0, "request_canary": { "foreign_marker_count": 0, "marker": "ANVIL_REQ_0_00088", "passed": true }, "request_index": 88, "time_to_first_output_ms": 1155.413700034842, "tpot_ms": 7.365560469603979, "ttft_ms": 1155.4143000394106, "visible_content_truncated": false, "visible_generation_ms": 3763.8007999630645 }, { "content_chunks": 45, "decode_tok_s": 244.72445200034235, "e2e_ms": 5025.534400017932, "effective_prefill_tok_s": 1252.7780267562327, "generation_ms": 2088.0627000005916, "mean_inter_token_latency_ms": 4.086228375735013, "output_token_source": "usage", "output_tokens": 512, "planned_context_tokens": 4096, "prompt_tokens": 3680, "reasoning_chunks": 0, "request_canary": { "foreign_marker_count": 0, "marker": "ANVIL_REQ_0_00089", "passed": true }, "request_index": 89, "time_to_first_output_ms": 2937.4717000173405, "tpot_ms": 4.086228375735013, "ttft_ms": 2937.4726000241935, "visible_content_truncated": false, "visible_generation_ms": 2088.0617999937385 }, { "content_chunks": 45, "decode_tok_s": 278.0519754794245, "e2e_ms": 5024.888799991459, "effective_prefill_tok_s": 1154.9674157042243, "generation_ms": 1837.7858999883756, "mean_inter_token_latency_ms": 3.5964499021298937, "output_token_source": "usage", "output_tokens": 512, "planned_context_tokens": 4096, "prompt_tokens": 3681, "reasoning_chunks": 0, "request_canary": { "foreign_marker_count": 0, "marker": "ANVIL_REQ_0_00090", "passed": true }, "request_index": 90, "time_to_first_output_ms": 3187.102900003083, "tpot_ms": 3.5964499021298937, "ttft_ms": 3187.103300006129, "visible_content_truncated": false, "visible_generation_ms": 1837.7854999853298 }, { "content_chunks": 44, "decode_tok_s": 210.2028515119642, "e2e_ms": 4879.787800018676, "effective_prefill_tok_s": 1503.591939011568, "generation_ms": 2430.9850999852642, "mean_inter_token_latency_ms": 4.757309393317542, "output_token_source": "usage", "output_tokens": 512, "planned_context_tokens": 4096, "prompt_tokens": 3682, "reasoning_chunks": 0, "request_canary": { "foreign_marker_count": 0, "marker": "ANVIL_REQ_0_00091", "passed": true }, "request_index": 91, "time_to_first_output_ms": 2448.8027000334114, "tpot_ms": 4.757309393317542, "ttft_ms": 2448.803100036457, "visible_content_truncated": false, "visible_generation_ms": 2430.9846999822184 }, { "content_chunks": 45, "decode_tok_s": 119.59990904850136, "e2e_ms": 9785.917600034736, "effective_prefill_tok_s": 668.0162299399237, "generation_ms": 4272.578499978408, "mean_inter_token_latency_ms": 8.361210371777707, "output_token_source": "usage", "output_tokens": 512, "planned_context_tokens": 4096, "prompt_tokens": 3683, "reasoning_chunks": 0, "request_canary": { "foreign_marker_count": 0, "marker": "ANVIL_REQ_0_00092", "passed": true }, "request_index": 92, "time_to_first_output_ms": 5513.339100056328, "tpot_ms": 8.361210371777707, "ttft_ms": 5513.339600060135, "visible_content_truncated": false, "visible_generation_ms": 4272.577999974601 }, { "content_chunks": 45, "decode_tok_s": 135.1014437327018, "e2e_ms": 4944.186600041576, "effective_prefill_tok_s": 3170.82264777238, "generation_ms": 3782.3430000571534, "mean_inter_token_latency_ms": 7.401845401286015, "output_token_source": "usage", "output_tokens": 512, "planned_context_tokens": 4096, "prompt_tokens": 3684, "reasoning_chunks": 0, "request_canary": { "foreign_marker_count": 0, "marker": "ANVIL_REQ_0_00093", "passed": true }, "request_index": 93, "time_to_first_output_ms": 1161.8435999844223, "tpot_ms": 7.401845401286015, "ttft_ms": 1161.8440999882296, "visible_content_truncated": false, "visible_generation_ms": 3782.342500053346 }, { "content_chunks": 45, "decode_tok_s": 144.76087995782606, "e2e_ms": 4943.787299911492, "effective_prefill_tok_s": 2606.3988967970195, "generation_ms": 3529.959199950099, "mean_inter_token_latency_ms": 6.907943639824069, "output_token_source": "usage", "output_tokens": 512, "planned_context_tokens": 4096, "prompt_tokens": 3685, "reasoning_chunks": 0, "request_canary": { "foreign_marker_count": 0, "marker": "ANVIL_REQ_0_00094", "passed": true }, "request_index": 94, "time_to_first_output_ms": 1413.8280999613926, "tpot_ms": 6.907943639824069, "ttft_ms": 1413.8286999659613, "visible_content_truncated": false, "visible_generation_ms": 3529.9585999455303 }, { "content_chunks": 45, "decode_tok_s": 208.9499179008372, "e2e_ms": 4943.565399968065, "effective_prefill_tok_s": 1475.5785150209026, "generation_ms": 2445.562099921517, "mean_inter_token_latency_ms": 4.785835811979485, "output_token_source": "usage", "output_tokens": 512, "planned_context_tokens": 4096, "prompt_tokens": 3686, "reasoning_chunks": 0, "request_canary": { "foreign_marker_count": 0, "marker": "ANVIL_REQ_0_00095", "passed": true }, "request_index": 95, "time_to_first_output_ms": 2498.0033000465482, "tpot_ms": 4.785835811979485, "ttft_ms": 2498.003900051117, "visible_content_truncated": false, "visible_generation_ms": 2445.5614999169484 }, { "content_chunks": 44, "decode_tok_s": 176.2845225317116, "e2e_ms": 4905.507700052112, "effective_prefill_tok_s": 1836.7690364763944, "generation_ms": 2898.7229999620467, "mean_inter_token_latency_ms": 5.672647749436491, "output_token_source": "usage", "output_tokens": 512, "planned_context_tokens": 4096, "prompt_tokens": 3686, "reasoning_chunks": 0, "request_canary": { "foreign_marker_count": 0, "marker": "ANVIL_REQ_0_00096", "passed": true }, "request_index": 96, "time_to_first_output_ms": 2006.7847000900656, "tpot_ms": 5.672647749436491, "ttft_ms": 2006.7851000931114, "visible_content_truncated": false, "visible_generation_ms": 2898.722599959001 }, { "content_chunks": 45, "decode_tok_s": 245.88554330992207, "e2e_ms": 4943.074699956924, "effective_prefill_tok_s": 1286.9684928568374, "generation_ms": 2078.2027000095695, "mean_inter_token_latency_ms": 4.066932876731056, "output_token_source": "usage", "output_tokens": 512, "planned_context_tokens": 4096, "prompt_tokens": 3687, "reasoning_chunks": 0, "request_canary": { "foreign_marker_count": 0, "marker": "ANVIL_REQ_0_00097", "passed": true }, "request_index": 97, "time_to_first_output_ms": 2864.871999947354, "tpot_ms": 4.066932876731056, "ttft_ms": 2864.872599951923, "visible_content_truncated": false, "visible_generation_ms": 2078.202100005001 }, { "content_chunks": 45, "decode_tok_s": 294.73728410593014, "e2e_ms": 4798.546400037594, "effective_prefill_tok_s": 1203.6678424744116, "generation_ms": 1733.7473999941722, "mean_inter_token_latency_ms": 3.392852054783116, "output_token_source": "usage", "output_tokens": 512, "planned_context_tokens": 4096, "prompt_tokens": 3689, "reasoning_chunks": 0, "request_canary": { "foreign_marker_count": 0, "marker": "ANVIL_REQ_0_00098", "passed": true }, "request_index": 98, "time_to_first_output_ms": 3064.799000043422, "tpot_ms": 3.392852054783116, "ttft_ms": 3064.799400046468, "visible_content_truncated": false, "visible_generation_ms": 1733.7469999911264 }, { "content_chunks": 44, "decode_tok_s": 362.26567046056687, "e2e_ms": 6715.234400006011, "effective_prefill_tok_s": 695.4253154543634, "generation_ms": 1410.5670000426471, "mean_inter_token_latency_ms": 2.7604050881460807, "output_token_source": "usage", "output_tokens": 512, "planned_context_tokens": 4096, "prompt_tokens": 3689, "reasoning_chunks": 0, "request_canary": { "foreign_marker_count": 0, "marker": "ANVIL_REQ_0_00099", "passed": true }, "request_index": 99, "time_to_first_output_ms": 5304.667399963364, "tpot_ms": 2.7604050881460807, "ttft_ms": 5304.66779996641, "visible_content_truncated": false, "visible_generation_ms": 1410.5666000396013 } ], "requests": 100, "response_words": 256, "run_id": "benchmark-20260905T035825Z", "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": "2026-09-05T03:58:25Z", "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", "distributions": "request-level mean, sample standard deviation, normal-approximation 95% confidence interval for the mean, min, p25, p50, p75, p90, p95, p99, and max using the nearest-rank percentile method", "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", "request_canaries": "when enabled, each response must contain its own synthetic marker and no marker assigned to another request; failures are excluded from performance metrics", "time_to_first_output": "request start through the first non-empty streamed reasoning or content delta", "tpot": "alias of mean_inter_token_latency for capacity-v3: client-observed generation time divided by completion tokens after the first token", "ttft": "request start through the first non-empty streamed content delta" }, "wall_clock_ms": 66989.42210001405 }