{ "base_url": "http://127.0.0.1:39261/v1", "completed": 100, "concurrency": 8, "context_distribution": { "4096": 100 }, "context_seed": 0, "context_tokens": 4096, "engine": "sglang", "failed": 0, "failures": [], "finished_at": "2026-09-05T04:29:35Z", "gpu": "NVIDIA RTX PRO 6000 Max-Q", "manifest": null, "max_context_tokens": 262144, "max_tokens": 512, "measurement_protocol": "capacity-v3", "metrics": { "completion_tokens_max": 512.0, "completion_tokens_mean": 494.39, "completion_tokens_mean_ci95_high": 503.36436663680047, "completion_tokens_mean_ci95_low": 485.4156333631995, "completion_tokens_min": 267.0, "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": 45.78758488163499, "content_chunks": 6284, "content_chunks_s": null, "decode_tok_s_max": 281.94148771362916, "decode_tok_s_mean": 136.61684377839106, "decode_tok_s_mean_ci95_high": 143.00011141901408, "decode_tok_s_mean_ci95_low": 130.23357613776804, "decode_tok_s_min": 101.74039184851588, "decode_tok_s_p25": 119.65952066074097, "decode_tok_s_p50": 122.47746456595466, "decode_tok_s_p75": 136.8470391283095, "decode_tok_s_p90": 171.23095453976856, "decode_tok_s_p95": 209.66872692198783, "decode_tok_s_p99": 268.9616408534496, "decode_tok_s_samples": 100, "decode_tok_s_stddev": 32.567692043995045, "e2e_max_ms": 8453.391100047156, "e2e_mean_ci95_high_ms": 5400.913602388687, "e2e_mean_ci95_low_ms": 5018.036827612445, "e2e_mean_ms": 5209.475215000566, "e2e_min_ms": 2712.8796000033617, "e2e_p25_ms": 4652.132199960761, "e2e_p50_ms": 5186.333800083958, "e2e_p75_ms": 5532.056399970315, "e2e_p90_ms": 6207.887399941683, "e2e_p95_ms": 7583.163899951614, "e2e_p99_ms": 8011.573199997656, "e2e_samples": 100, "e2e_stddev_ms": 976.7264662659222, "effective_prefill_tok_s_max": 5666.072458556901, "effective_prefill_tok_s_mean": 2944.4499878024744, "effective_prefill_tok_s_mean_ci95_high": 3138.8889891568915, "effective_prefill_tok_s_mean_ci95_low": 2750.0109864480573, "effective_prefill_tok_s_min": 819.6043198265872, "effective_prefill_tok_s_p25": 2316.4073963457754, "effective_prefill_tok_s_p50": 2909.7331285700297, "effective_prefill_tok_s_p75": 3736.645391262173, "effective_prefill_tok_s_p90": 4416.105959436682, "effective_prefill_tok_s_p95": 4443.711868341752, "effective_prefill_tok_s_p99": 4505.780049379425, "effective_prefill_tok_s_samples": 100, "effective_prefill_tok_s_stddev": 992.0357211960068, "generation_max_ms": 5022.587300045416, "generation_mean_ci95_high_ms": 3910.235864392334, "generation_mean_ci95_low_ms": 3619.6539536068763, "generation_mean_ms": 3764.944908999605, "generation_min_ms": 1130.2726999856532, "generation_p25_ms": 3418.029800057411, "generation_p50_ms": 4041.026399936527, "generation_p75_ms": 4253.509199945256, "generation_p90_ms": 4358.298099949025, "generation_p95_ms": 4479.593999916688, "generation_p99_ms": 4726.846999954432, "generation_samples": 100, "generation_stddev_ms": 741.2803846567801, "mean_inter_token_latency_ms_max": 9.828937964863828, "mean_inter_token_latency_ms_mean": 7.610339859626578, "mean_inter_token_latency_ms_mean_ci95_high": 7.85991035230111, "mean_inter_token_latency_ms_mean_ci95_low": 7.360769366952045, "mean_inter_token_latency_ms_min": 3.54683522495884, "mean_inter_token_latency_ms_p25": 7.2415885594436675, "mean_inter_token_latency_ms_p50": 8.132569710433666, "mean_inter_token_latency_ms_p75": 8.353640508761451, "mean_inter_token_latency_ms_p90": 8.616821721939775, "mean_inter_token_latency_ms_p95": 8.945890606851217, "mean_inter_token_latency_ms_p99": 9.44189773238942, "mean_inter_token_latency_ms_samples": 100, "mean_inter_token_latency_ms_stddev": 1.273318840176187, "output_token_sources": { "usage": 100 }, "output_tokens": 49439, "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": 746.6756478980565, "time_to_first_output_max_ms": 4419.205600046553, "time_to_first_output_mean_ci95_high_ms": 1581.80928421037, "time_to_first_output_mean_ci95_low_ms": 1307.251327791552, "time_to_first_output_mean_ms": 1444.530306000961, "time_to_first_output_min_ms": 637.6550999702886, "time_to_first_output_p25_ms": 978.4538999665529, "time_to_first_output_p50_ms": 1238.9612001134083, "time_to_first_output_p75_ms": 1565.077199949883, "time_to_first_output_p90_ms": 2135.990099981427, "time_to_first_output_p95_ms": 3263.697799993679, "time_to_first_output_p99_ms": 4030.213400023058, "time_to_first_output_samples": 100, "time_to_first_output_stddev_ms": 700.4029500480045, "tpot_max_ms": 9.828937964863828, "tpot_mean_ci95_high_ms": 7.85991035230111, "tpot_mean_ci95_low_ms": 7.360769366952045, "tpot_mean_ms": 7.610339859626578, "tpot_min_ms": 3.54683522495884, "tpot_p25_ms": 7.2415885594436675, "tpot_p50_ms": 8.132569710433666, "tpot_p75_ms": 8.353640508761451, "tpot_p90_ms": 8.616821721939775, "tpot_p95_ms": 8.945890606851217, "tpot_p99_ms": 9.44189773238942, "tpot_samples": 100, "tpot_stddev_ms": 1.273318840176187, "ttft_max_ms": 4419.205800048076, "ttft_mean_ci95_high_ms": 1581.8097016016761, "ttft_mean_ci95_low_ms": 1307.251740392596, "ttft_mean_ms": 1444.530720997136, "ttft_min_ms": 637.6557999756187, "ttft_p25_ms": 978.4543999703601, "ttft_p50_ms": 1238.9617000008002, "ttft_p75_ms": 1565.0775999529287, "ttft_p90_ms": 2135.9903999837115, "ttft_p95_ms": 3263.699200004339, "ttft_p99_ms": 4030.2139000268653, "ttft_samples": 100, "ttft_stddev_ms": 700.4029622680612 }, "model": "qwen38-27b-radixark-nvfp4-sglang-pro6000-c8-dflash2", "prompt_cache_mode": "unique", "request_canaries": true, "request_timings": [ { "content_chunks": 65, "decode_tok_s": 120.96754625287439, "e2e_ms": 4861.9286000030115, "effective_prefill_tok_s": 5666.072458556901, "generation_ms": 4224.273500032723, "mean_inter_token_latency_ms": 8.26668003920298, "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": 637.6550999702886, "tpot_ms": 8.26668003920298, "ttft_ms": 637.6557999756187, "visible_content_truncated": false, "visible_generation_ms": 4224.272800027393 }, { "content_chunks": 65, "decode_tok_s": 134.23920012322074, "e2e_ms": 4862.394799944013, "effective_prefill_tok_s": 3424.08305851392, "generation_ms": 3806.637699948624, "mean_inter_token_latency_ms": 7.449388845300634, "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": 1055.7570999953896, "tpot_ms": 7.449388845300634, "ttft_ms": 1055.758100003004, "visible_content_truncated": false, "visible_generation_ms": 3806.6366999410093 }, { "content_chunks": 48, "decode_tok_s": 148.27696403722842, "e2e_ms": 3984.295100090094, "effective_prefill_tok_s": 2462.667807802599, "generation_ms": 2515.5626999912784, "mean_inter_token_latency_ms": 6.744135924909593, "output_token_source": "usage", "output_tokens": 374, "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": 1468.7324000988156, "tpot_ms": 6.744135924909593, "ttft_ms": 1468.7328001018614, "visible_content_truncated": false, "visible_generation_ms": 2515.5622999882326 }, { "content_chunks": 65, "decode_tok_s": 171.23095453976856, "e2e_ms": 4862.117299926467, "effective_prefill_tok_s": 1926.1452949379334, "generation_ms": 2984.273499925621, "mean_inter_token_latency_ms": 5.840065557584385, "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": 1877.8438000008464, "tpot_ms": 5.840065557584385, "ttft_ms": 1877.8441000031307, "visible_content_truncated": false, "visible_generation_ms": 2984.2731999233365 }, { "content_chunks": 65, "decode_tok_s": 183.74094697905133, "e2e_ms": 4860.381199978292, "effective_prefill_tok_s": 1740.4964517910262, "generation_ms": 2781.089400057681, "mean_inter_token_latency_ms": 5.4424450098976145, "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": 2079.291799920611, "tpot_ms": 5.4424450098976145, "ttft_ms": 2079.2920999228954, "visible_content_truncated": false, "visible_generation_ms": 2781.0891000553966 }, { "content_chunks": 65, "decode_tok_s": 215.4510805608496, "e2e_ms": 4861.547700013034, "effective_prefill_tok_s": 1453.9439555549177, "generation_ms": 2371.768100070767, "mean_inter_token_latency_ms": 4.64142485336745, "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": 2489.779599942267, "tpot_ms": 4.64142485336745, "ttft_ms": 2489.780099946074, "visible_content_truncated": false, "visible_generation_ms": 2371.76760006696 }, { "content_chunks": 65, "decode_tok_s": 226.85446310110433, "e2e_ms": 4860.17619993072, "effective_prefill_tok_s": 1389.0003672739176, "generation_ms": 2252.545499941334, "mean_inter_token_latency_ms": 4.408112524347033, "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": 2607.630699989386, "tpot_ms": 4.408112524347033, "ttft_ms": 2607.6309999916703, "visible_content_truncated": false, "visible_generation_ms": 2252.54519993905 }, { "content_chunks": 63, "decode_tok_s": 123.19711128799746, "e2e_ms": 8453.391100047156, "effective_prefill_tok_s": 819.6043198265872, "generation_ms": 4034.1855000006035, "mean_inter_token_latency_ms": 8.117073440645077, "output_token_source": "usage", "output_tokens": 498, "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": 4419.205600046553, "tpot_ms": 8.117073440645077, "ttft_ms": 4419.205800048076, "visible_content_truncated": false, "visible_generation_ms": 4034.1852999990806 }, { "content_chunks": 66, "decode_tok_s": 119.93423942024656, "e2e_ms": 5791.18830000516, "effective_prefill_tok_s": 2366.5158008733288, "generation_ms": 4260.668200091459, "mean_inter_token_latency_ms": 8.33790254421029, "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": 1530.5200999137014, "tpot_ms": 8.33790254421029, "ttft_ms": 1530.5205999175087, "visible_content_truncated": false, "visible_generation_ms": 4260.667700087652 }, { "content_chunks": 65, "decode_tok_s": 182.7125011688726, "e2e_ms": 4885.401299921796, "effective_prefill_tok_s": 1735.085565483376, "generation_ms": 2796.7434999300167, "mean_inter_token_latency_ms": 5.473079256223125, "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": 2088.657799991779, "tpot_ms": 5.473079256223125, "ttft_ms": 2088.658199994825, "visible_content_truncated": false, "visible_generation_ms": 2796.743099926971 }, { "content_chunks": 49, "decode_tok_s": 148.89810759466422, "e2e_ms": 4061.6550999693573, "effective_prefill_tok_s": 2456.6490534869536, "generation_ms": 2585.6607999885455, "mean_inter_token_latency_ms": 6.716002077892326, "output_token_source": "usage", "output_tokens": 386, "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": 1475.9942999808118, "tpot_ms": 6.716002077892326, "ttft_ms": 1475.9946999838576, "visible_content_truncated": false, "visible_generation_ms": 2585.6603999854997 }, { "content_chunks": 65, "decode_tok_s": 133.72909118709046, "e2e_ms": 4885.512700071558, "effective_prefill_tok_s": 3407.6989002337523, "generation_ms": 3821.158100035973, "mean_inter_token_latency_ms": 7.477804501048871, "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": 1064.3546000355855, "tpot_ms": 7.477804501048871, "ttft_ms": 1064.3551000393927, "visible_content_truncated": false, "visible_generation_ms": 3821.1576000321656 }, { "content_chunks": 65, "decode_tok_s": 193.3942650191221, "e2e_ms": 4885.369400028139, "effective_prefill_tok_s": 1616.9596103315025, "generation_ms": 2642.270699958317, "mean_inter_token_latency_ms": 5.170784148646413, "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": 2243.0987000698224, "tpot_ms": 5.170784148646413, "ttft_ms": 2243.0994000751525, "visible_content_truncated": false, "visible_generation_ms": 2642.269999952987 }, { "content_chunks": 65, "decode_tok_s": 170.30711172024016, "e2e_ms": 4885.424599982798, "effective_prefill_tok_s": 1924.70652064178, "generation_ms": 3000.4618999082595, "mean_inter_token_latency_ms": 5.871745400994637, "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": 1884.9627000745386, "tpot_ms": 5.871745400994637, "ttft_ms": 1884.9631999619305, "visible_content_truncated": false, "visible_generation_ms": 3000.4614000208676 }, { "content_chunks": 61, "decode_tok_s": 120.05948319542729, "e2e_ms": 8011.573199997656, "effective_prefill_tok_s": 900.4485965877731, "generation_ms": 3981.359799974598, "mean_inter_token_latency_ms": 8.329204602457319, "output_token_source": "usage", "output_tokens": 479, "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": 4030.213400023058, "tpot_ms": 8.329204602457319, "ttft_ms": 4030.2139000268653, "visible_content_truncated": false, "visible_generation_ms": 3981.3592999707907 }, { "content_chunks": 65, "decode_tok_s": 114.0728378530518, "e2e_ms": 5386.716199922375, "effective_prefill_tok_s": 4001.665927674617, "generation_ms": 4479.593999916688, "mean_inter_token_latency_ms": 8.76632876696025, "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": 907.1222000056878, "tpot_ms": 8.76632876696025, "ttft_ms": 907.1228998946026, "visible_content_truncated": false, "visible_generation_ms": 4479.593300027773 }, { "content_chunks": 65, "decode_tok_s": 119.95668694821732, "e2e_ms": 5743.201300036162, "effective_prefill_tok_s": 2446.5216920603934, "generation_ms": 4259.87090007402, "mean_inter_token_latency_ms": 8.336342270203561, "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": 1483.330399962142, "tpot_ms": 8.336342270203561, "ttft_ms": 1483.330799965188, "visible_content_truncated": false, "visible_generation_ms": 4259.870500070974 }, { "content_chunks": 65, "decode_tok_s": 132.7990773623138, "e2e_ms": 4920.220699976198, "effective_prefill_tok_s": 3385.2390892147273, "generation_ms": 3847.91829995811, "mean_inter_token_latency_ms": 7.530172798352465, "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": 1072.3024000180885, "tpot_ms": 7.530172798352465, "ttft_ms": 1072.3029000218958, "visible_content_truncated": false, "visible_generation_ms": 3847.9177999543026 }, { "content_chunks": 65, "decode_tok_s": 168.10951897794303, "e2e_ms": 4919.410299975425, "effective_prefill_tok_s": 1931.6653306478972, "generation_ms": 3039.6850999677554, "mean_inter_token_latency_ms": 5.948503131052359, "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": 1879.7252000076696, "tpot_ms": 5.948503131052359, "ttft_ms": 1879.7256000107154, "visible_content_truncated": false, "visible_generation_ms": 3039.6846999647096 }, { "content_chunks": 65, "decode_tok_s": 148.80088761102806, "e2e_ms": 4919.619500054978, "effective_prefill_tok_s": 2444.9676949421155, "generation_ms": 3434.1192999854684, "mean_inter_token_latency_ms": 6.720390019541034, "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": 1485.5002000695094, "tpot_ms": 6.720390019541034, "ttft_ms": 1485.5005000717938, "visible_content_truncated": false, "visible_generation_ms": 3434.118999983184 }, { "content_chunks": 46, "decode_tok_s": 209.66872692198783, "e2e_ms": 3595.2039000112563, "effective_prefill_tok_s": 1924.514838725993, "generation_ms": 1707.455400028266, "mean_inter_token_latency_ms": 4.769428491699067, "output_token_source": "usage", "output_tokens": 359, "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": 1887.7484999829903, "tpot_ms": 4.769428491699067, "ttft_ms": 1887.7489999867976, "visible_content_truncated": false, "visible_generation_ms": 1707.4549000244588 }, { "content_chunks": 65, "decode_tok_s": 122.47746456595466, "e2e_ms": 7709.671900025569, "effective_prefill_tok_s": 1027.2861795899871, "generation_ms": 4172.196100000292, "mean_inter_token_latency_ms": 8.164767318982959, "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": 3537.475800025277, "tpot_ms": 8.164767318982959, "ttft_ms": 3537.4760000267997, "visible_content_truncated": false, "visible_generation_ms": 4172.195899998769 }, { "content_chunks": 65, "decode_tok_s": 122.29806537038618, "e2e_ms": 5086.212299996987, "effective_prefill_tok_s": 4003.7625456635374, "generation_ms": 4178.316300036386, "mean_inter_token_latency_ms": 8.176744227077076, "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": 907.8959999606013, "tpot_ms": 8.176744227077076, "ttft_ms": 907.8962999628857, "visible_content_truncated": false, "visible_generation_ms": 4178.316000034101 }, { "content_chunks": 61, "decode_tok_s": 130.48242287774332, "e2e_ms": 4616.902099922299, "effective_prefill_tok_s": 3875.317520881388, "generation_ms": 3678.6563999485224, "mean_inter_token_latency_ms": 7.663867499892755, "output_token_source": "usage", "output_tokens": 481, "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": 938.2456999737769, "tpot_ms": 7.663867499892755, "ttft_ms": 938.2462999783456, "visible_content_truncated": false, "visible_generation_ms": 3678.6557999439538 }, { "content_chunks": 65, "decode_tok_s": 129.3459446395055, "e2e_ms": 5436.544899945147, "effective_prefill_tok_s": 2447.0031345914454, "generation_ms": 3950.6456999806687, "mean_inter_token_latency_ms": 7.731204892330076, "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": 1485.8991999644786, "tpot_ms": 7.731204892330076, "ttft_ms": 1485.8995999675244, "visible_content_truncated": false, "visible_generation_ms": 3950.645299977623 }, { "content_chunks": 65, "decode_tok_s": 163.72132746363056, "e2e_ms": 4612.780399969779, "effective_prefill_tok_s": 2438.2833412448786, "generation_ms": 3121.1571999592707, "mean_inter_token_latency_ms": 6.107939725947692, "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": 1491.6232000105083, "tpot_ms": 6.107939725947692, "ttft_ms": 1491.6237000143155, "visible_content_truncated": false, "visible_generation_ms": 3121.1566999554634 }, { "content_chunks": 65, "decode_tok_s": 144.54098267299497, "e2e_ms": 4611.435199971311, "effective_prefill_tok_s": 3381.6380677389348, "generation_ms": 3535.329499980435, "mean_inter_token_latency_ms": 6.918453033229814, "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": 1076.105699990876, "tpot_ms": 6.918453033229814, "ttft_ms": 1076.1061999946833, "visible_content_truncated": false, "visible_generation_ms": 3535.3289999766275 }, { "content_chunks": 63, "decode_tok_s": 166.04072875548437, "e2e_ms": 4580.635400023311, "effective_prefill_tok_s": 2292.4366787163262, "generation_ms": 2993.241499993019, "mean_inter_token_latency_ms": 6.022618712259595, "output_token_source": "usage", "output_tokens": 498, "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": 1587.3939000302926, "tpot_ms": 6.022618712259595, "ttft_ms": 1587.3945000348613, "visible_content_truncated": false, "visible_generation_ms": 2993.24089998845 }, { "content_chunks": 65, "decode_tok_s": 116.97784643034248, "e2e_ms": 7632.046300102957, "effective_prefill_tok_s": 1115.299339297606, "generation_ms": 4368.348500109278, "mean_inter_token_latency_ms": 8.548627201779409, "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": 3263.697799993679, "tpot_ms": 8.548627201779409, "ttft_ms": 3263.699200004339, "visible_content_truncated": false, "visible_generation_ms": 4368.347100098617 }, { "content_chunks": 60, "decode_tok_s": 138.0912477685455, "e2e_ms": 4561.845600022934, "effective_prefill_tok_s": 3184.0791149324714, "generation_ms": 3418.029800057411, "mean_inter_token_latency_ms": 7.2415885594436675, "output_token_source": "usage", "output_tokens": 473, "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": 1143.8157999655232, "tpot_ms": 7.2415885594436675, "ttft_ms": 1143.816199968569, "visible_content_truncated": false, "visible_generation_ms": 3418.0294000543654 }, { "content_chunks": 65, "decode_tok_s": 122.39639689661648, "e2e_ms": 4986.829200061038, "effective_prefill_tok_s": 4485.941524998864, "generation_ms": 4174.959500087425, "mean_inter_token_latency_ms": 8.170175146942123, "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": 811.869699973613, "tpot_ms": 8.170175146942123, "ttft_ms": 811.8700999766588, "visible_content_truncated": false, "visible_generation_ms": 4174.959100084379 }, { "content_chunks": 65, "decode_tok_s": 129.4712326457757, "e2e_ms": 5905.501100001857, "effective_prefill_tok_s": 1859.9275919664658, "generation_ms": 3946.822699974291, "mean_inter_token_latency_ms": 7.723723483315638, "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": 1958.6784000275657, "tpot_ms": 7.723723483315638, "ttft_ms": 1958.679600036703, "visible_content_truncated": false, "visible_generation_ms": 3946.8214999651536 }, { "content_chunks": 65, "decode_tok_s": 134.49407422120186, "e2e_ms": 4618.058199994266, "effective_prefill_tok_s": 4451.316051476984, "generation_ms": 3799.4238999672234, "mean_inter_token_latency_ms": 7.435271819896719, "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": 818.6343000270426, "tpot_ms": 7.435271819896719, "ttft_ms": 818.6347000300884, "visible_content_truncated": false, "visible_generation_ms": 3799.4234999641776 }, { "content_chunks": 65, "decode_tok_s": 167.86748862783634, "e2e_ms": 4617.625099956058, "effective_prefill_tok_s": 2316.4073963457754, "generation_ms": 3044.0676999278367, "mean_inter_token_latency_ms": 5.957079647608291, "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": 1573.5574000282213, "tpot_ms": 5.957079647608291, "ttft_ms": 1573.5579000320286, "visible_content_truncated": false, "visible_generation_ms": 3044.0671999240294 }, { "content_chunks": 65, "decode_tok_s": 160.18067627866222, "e2e_ms": 4617.630900000222, "effective_prefill_tok_s": 2554.145467096698, "generation_ms": 3190.147600020282, "mean_inter_token_latency_ms": 6.242950293581765, "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": 1427.48329997994, "tpot_ms": 6.242950293581765, "ttft_ms": 1427.4836999829859, "visible_content_truncated": false, "visible_generation_ms": 3190.147200017236 }, { "content_chunks": 65, "decode_tok_s": 123.18046364743925, "e2e_ms": 7583.163899951614, "effective_prefill_tok_s": 1061.4948479356365, "generation_ms": 4148.3850999502465, "mean_inter_token_latency_ms": 8.118170450000482, "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": 3434.778800001368, "tpot_ms": 8.118170450000482, "ttft_ms": 3434.7791000036523, "visible_content_truncated": false, "visible_generation_ms": 4148.384799947962 }, { "content_chunks": 65, "decode_tok_s": 127.55930840044316, "e2e_ms": 4815.384500077926, "effective_prefill_tok_s": 4505.780049379425, "generation_ms": 4005.979700013995, "mean_inter_token_latency_ms": 7.839490606681008, "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": 809.4048000639305, "tpot_ms": 7.839490606681008, "ttft_ms": 809.4054000684991, "visible_content_truncated": false, "visible_generation_ms": 4005.9791000094265 }, { "content_chunks": 65, "decode_tok_s": 111.78316882548835, "e2e_ms": 5661.043000058271, "effective_prefill_tok_s": 3348.6498812123946, "generation_ms": 4571.350100100972, "mean_inter_token_latency_ms": 8.945890606851217, "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": 1089.6928999572992, "tpot_ms": 8.945890606851217, "ttft_ms": 1089.6931999595836, "visible_content_truncated": false, "visible_generation_ms": 4571.349800098687 }, { "content_chunks": 65, "decode_tok_s": 108.10588961413943, "e2e_ms": 6314.573099953122, "effective_prefill_tok_s": 2298.2553476969424, "generation_ms": 4726.846999954432, "mean_inter_token_latency_ms": 9.25018982378558, "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": 1587.7260999986902, "tpot_ms": 9.25018982378558, "ttft_ms": 1587.7264000009745, "visible_content_truncated": false, "visible_generation_ms": 4726.846699952148 }, { "content_chunks": 66, "decode_tok_s": 120.9675090254176, "e2e_ms": 5425.430800067261, "effective_prefill_tok_s": 3038.7393476993234, "generation_ms": 4224.274800042622, "mean_inter_token_latency_ms": 8.266682583253663, "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": 1201.156000024639, "tpot_ms": 8.266682583253663, "ttft_ms": 1201.1564000276849, "visible_content_truncated": false, "visible_generation_ms": 4224.274400039576 }, { "content_chunks": 52, "decode_tok_s": 113.93199331586901, "e2e_ms": 4629.221900017001, "effective_prefill_tok_s": 3370.4227540321667, "generation_ms": 3545.97500001546, "mean_inter_token_latency_ms": 8.777165841622425, "output_token_source": "usage", "output_tokens": 405, "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": 1083.2469000015408, "tpot_ms": 8.777165841622425, "ttft_ms": 1083.2473000045866, "visible_content_truncated": false, "visible_generation_ms": 3545.974600012414 }, { "content_chunks": 65, "decode_tok_s": 117.24760176592251, "e2e_ms": 7740.535199991427, "effective_prefill_tok_s": 1079.7587194446585, "generation_ms": 4358.298099949025, "mean_inter_token_latency_ms": 8.528959099704549, "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": 3382.2371000424027, "tpot_ms": 8.528959099704549, "ttft_ms": 3382.237400044687, "visible_content_truncated": false, "visible_generation_ms": 4358.29779994674 }, { "content_chunks": 39, "decode_tok_s": 268.9616408534496, "e2e_ms": 2712.8796000033617, "effective_prefill_tok_s": 2308.216904626869, "generation_ms": 1130.2726999856532, "mean_inter_token_latency_ms": 3.7180023025843854, "output_token_source": "usage", "output_tokens": 305, "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": 1582.6069000177085, "tpot_ms": 3.7180023025843854, "ttft_ms": 1582.607299904339, "visible_content_truncated": false, "visible_generation_ms": 1130.2723000990227 }, { "content_chunks": 63, "decode_tok_s": 119.02323735840599, "e2e_ms": 4997.715699952096, "effective_prefill_tok_s": 4443.711868341752, "generation_ms": 4175.655200029723, "mean_inter_token_latency_ms": 8.401720724405882, "output_token_source": "usage", "output_tokens": 498, "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": 822.0604999223724, "tpot_ms": 8.401720724405882, "ttft_ms": 822.0607999246567, "visible_content_truncated": false, "visible_generation_ms": 4175.654900027439 }, { "content_chunks": 65, "decode_tok_s": 121.65704996546279, "e2e_ms": 5742.99489997793, "effective_prefill_tok_s": 2368.631539748095, "generation_ms": 4200.332000036724, "mean_inter_token_latency_ms": 8.219827788721574, "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": 1542.662899941206, "tpot_ms": 8.219827788721574, "ttft_ms": 1542.6633999450132, "visible_content_truncated": false, "visible_generation_ms": 4200.331500032917 }, { "content_chunks": 55, "decode_tok_s": 136.8470391283095, "e2e_ms": 4119.766499963589, "effective_prefill_tok_s": 3736.808773474995, "generation_ms": 3142.1943999594077, "mean_inter_token_latency_ms": 7.307428837114902, "output_token_source": "usage", "output_tokens": 431, "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": 977.5721000041813, "tpot_ms": 7.307428837114902, "ttft_ms": 977.5723000057042, "visible_content_truncated": false, "visible_generation_ms": 3142.194199957885 }, { "content_chunks": 65, "decode_tok_s": 121.60490585085837, "e2e_ms": 5175.311899976805, "effective_prefill_tok_s": 3754.7057130101193, "generation_ms": 4202.13310001418, "mean_inter_token_latency_ms": 8.223352446211702, "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": 973.1787999626249, "tpot_ms": 8.223352446211702, "ttft_ms": 973.1792999664322, "visible_content_truncated": false, "visible_generation_ms": 4202.132600010373 }, { "content_chunks": 65, "decode_tok_s": 128.46616999109045, "e2e_ms": 5415.049400064163, "effective_prefill_tok_s": 2542.8766927249403, "generation_ms": 3977.7009000536054, "mean_inter_token_latency_ms": 7.784150489341694, "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": 1437.3485000105575, "tpot_ms": 7.784150489341694, "ttft_ms": 1437.3489000136033, "visible_content_truncated": false, "visible_generation_ms": 3977.7005000505596 }, { "content_chunks": 65, "decode_tok_s": 133.64592006514266, "e2e_ms": 4648.346800007857, "effective_prefill_tok_s": 4431.319817266284, "generation_ms": 3823.5361000988632, "mean_inter_token_latency_ms": 7.4824581215241945, "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": 824.8106999089941, "tpot_ms": 7.4824581215241945, "ttft_ms": 824.8110999120399, "visible_content_truncated": false, "visible_generation_ms": 3823.5357000958174 }, { "content_chunks": 65, "decode_tok_s": 121.5081823840922, "e2e_ms": 6341.468200087547, "effective_prefill_tok_s": 1711.1502530052835, "generation_ms": 4205.47810010612, "mean_inter_token_latency_ms": 8.229898434649941, "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": 2135.990099981427, "tpot_ms": 8.229898434649941, "ttft_ms": 2135.9903999837115, "visible_content_truncated": false, "visible_generation_ms": 4205.477800103836 }, { "content_chunks": 64, "decode_tok_s": 126.27842992089256, "e2e_ms": 5246.67539994698, "effective_prefill_tok_s": 2822.197685040376, "generation_ms": 3951.585399918258, "mean_inter_token_latency_ms": 7.919008817471459, "output_token_source": "usage", "output_tokens": 500, "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": 1295.090000028722, "tpot_ms": 7.919008817471459, "ttft_ms": 1295.0903999153525, "visible_content_truncated": false, "visible_generation_ms": 3951.5850000316277 }, { "content_chunks": 65, "decode_tok_s": 133.2964068391413, "e2e_ms": 4652.132199960761, "effective_prefill_tok_s": 4465.1010511241175, "generation_ms": 3833.5616999538615, "mean_inter_token_latency_ms": 7.502077690712058, "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": 818.5705000068992, "tpot_ms": 7.502077690712058, "ttft_ms": 818.5708000091836, "visible_content_truncated": false, "visible_generation_ms": 3833.561399951577 }, { "content_chunks": 65, "decode_tok_s": 121.04033480804438, "e2e_ms": 5630.883499979973, "effective_prefill_tok_s": 2593.76164493577, "generation_ms": 4221.733200014569, "mean_inter_token_latency_ms": 8.261708806290741, "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": 1409.150299965404, "tpot_ms": 8.261708806290741, "ttft_ms": 1409.1506999684498, "visible_content_truncated": false, "visible_generation_ms": 4221.732800011523 }, { "content_chunks": 65, "decode_tok_s": 110.94578765541942, "e2e_ms": 5581.840500002727, "effective_prefill_tok_s": 3745.949998773144, "generation_ms": 4605.85309995804, "mean_inter_token_latency_ms": 9.013411154516712, "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": 975.9874000446871, "tpot_ms": 9.013411154516712, "ttft_ms": 975.9877000469714, "visible_content_truncated": false, "visible_generation_ms": 4605.8527999557555 }, { "content_chunks": 65, "decode_tok_s": 116.53911221492869, "e2e_ms": 5829.845799948089, "effective_prefill_tok_s": 2530.7051276745096, "generation_ms": 4384.793999954127, "mean_inter_token_latency_ms": 8.580810176035474, "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": 1445.051799993962, "tpot_ms": 8.580810176035474, "ttft_ms": 1445.0520999962464, "visible_content_truncated": false, "visible_generation_ms": 4384.793699951842 }, { "content_chunks": 64, "decode_tok_s": 118.37620084773113, "e2e_ms": 6375.838000094518, "effective_prefill_tok_s": 1686.5604416236329, "generation_ms": 4206.92670007702, "mean_inter_token_latency_ms": 8.447643976058272, "output_token_source": "usage", "output_tokens": 499, "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": 2168.9113000174984, "tpot_ms": 8.447643976058272, "ttft_ms": 2168.9117000205442, "visible_content_truncated": false, "visible_generation_ms": 4206.926300073974 }, { "content_chunks": 48, "decode_tok_s": 145.1949714764822, "e2e_ms": 3387.6445000059903, "effective_prefill_tok_s": 4432.0769967487695, "generation_ms": 2562.0722000021487, "mean_inter_token_latency_ms": 6.88729086022083, "output_token_source": "usage", "output_tokens": 373, "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": 825.5723000038415, "tpot_ms": 6.88729086022083, "ttft_ms": 825.5727000068873, "visible_content_truncated": false, "visible_generation_ms": 2562.071799999103 }, { "content_chunks": 62, "decode_tok_s": 123.79753033850787, "e2e_ms": 5186.333800083958, "effective_prefill_tok_s": 2884.964181345793, "generation_ms": 3917.687200009823, "mean_inter_token_latency_ms": 8.077705567030563, "output_token_source": "usage", "output_tokens": 486, "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": 1268.6466000741348, "tpot_ms": 8.077705567030563, "ttft_ms": 1268.6470000771806, "visible_content_truncated": false, "visible_generation_ms": 3917.686800006777 }, { "content_chunks": 65, "decode_tok_s": 120.49249029755772, "e2e_ms": 5068.984900019132, "effective_prefill_tok_s": 4421.194829098721, "generation_ms": 4240.92820007354, "mean_inter_token_latency_ms": 8.299272407188925, "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": 828.0566999455914, "tpot_ms": 8.299272407188925, "ttft_ms": 828.0570999486372, "visible_content_truncated": false, "visible_generation_ms": 4240.927800070494 }, { "content_chunks": 65, "decode_tok_s": 116.05207027248153, "e2e_ms": 6085.142600000836, "effective_prefill_tok_s": 2176.644479759643, "generation_ms": 4403.195899911225, "mean_inter_token_latency_ms": 8.616821721939775, "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": 1681.946700089611, "tpot_ms": 8.616821721939775, "ttft_ms": 1681.9470000918955, "visible_content_truncated": false, "visible_generation_ms": 4403.1955999089405 }, { "content_chunks": 60, "decode_tok_s": 126.83915426912928, "e2e_ms": 4541.851700050756, "effective_prefill_tok_s": 4420.103965571651, "generation_ms": 3713.364400085993, "mean_inter_token_latency_ms": 7.8840008494394755, "output_token_source": "usage", "output_tokens": 472, "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": 828.4872999647632, "tpot_ms": 7.8840008494394755, "ttft_ms": 828.4875999670476, "visible_content_truncated": false, "visible_generation_ms": 3713.3641000837088 }, { "content_chunks": 65, "decode_tok_s": 120.44560063965898, "e2e_ms": 5580.27429995127, "effective_prefill_tok_s": 2738.292156420697, "generation_ms": 4242.579199955799, "mean_inter_token_latency_ms": 8.302503326723677, "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": 1337.6950999954715, "tpot_ms": 8.302503326723677, "ttft_ms": 1337.6954999985173, "visible_content_truncated": false, "visible_generation_ms": 4242.578799952753 }, { "content_chunks": 65, "decode_tok_s": 120.44114929652342, "e2e_ms": 5470.441200071946, "effective_prefill_tok_s": 2985.2443402066156, "generation_ms": 4242.735999985598, "mean_inter_token_latency_ms": 8.302810176097061, "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": 1227.7052000863478, "tpot_ms": 8.302810176097061, "ttft_ms": 1227.705500088632, "visible_content_truncated": false, "visible_generation_ms": 4242.735699983314 }, { "content_chunks": 65, "decode_tok_s": 114.97030346281021, "e2e_ms": 6207.887399941683, "effective_prefill_tok_s": 2078.534697206429, "generation_ms": 4444.62600001134, "mean_inter_token_latency_ms": 8.697898238769746, "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": 1763.261399930343, "tpot_ms": 8.697898238769746, "ttft_ms": 1763.2616999326274, "visible_content_truncated": false, "visible_generation_ms": 4444.625700009055 }, { "content_chunks": 57, "decode_tok_s": 122.96236437015126, "e2e_ms": 4632.617700030096, "effective_prefill_tok_s": 3736.645391262173, "generation_ms": 3651.523799984716, "mean_inter_token_latency_ms": 8.132569710433666, "output_token_source": "usage", "output_tokens": 450, "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": 981.09390004538, "tpot_ms": 8.132569710433666, "ttft_ms": 981.0942000476643, "visible_content_truncated": false, "visible_generation_ms": 3651.5234999824315 }, { "content_chunks": 65, "decode_tok_s": 120.42711609461503, "e2e_ms": 5172.031600028276, "effective_prefill_tok_s": 3948.099980907348, "generation_ms": 4243.230400024913, "mean_inter_token_latency_ms": 8.303777690851103, "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": 928.8012000033632, "tpot_ms": 8.303777690851103, "ttft_ms": 928.8017000071704, "visible_content_truncated": false, "visible_generation_ms": 4243.229900021106 }, { "content_chunks": 65, "decode_tok_s": 120.28868248166448, "e2e_ms": 5202.867900021374, "effective_prefill_tok_s": 3839.731733831037, "generation_ms": 4248.113699955866, "mean_inter_token_latency_ms": 8.313334050794257, "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": 954.7542000655085, "tpot_ms": 8.313334050794257, "ttft_ms": 954.7545000677928, "visible_content_truncated": false, "visible_generation_ms": 4248.113399953581 }, { "content_chunks": 64, "decode_tok_s": 117.918508177456, "e2e_ms": 5201.709499931894, "effective_prefill_tok_s": 3747.749383108751, "generation_ms": 4223.255599965341, "mean_inter_token_latency_ms": 8.480432931657312, "output_token_source": "usage", "output_tokens": 499, "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": 978.4538999665529, "tpot_ms": 8.480432931657312, "ttft_ms": 978.4543999703601, "visible_content_truncated": false, "visible_generation_ms": 4223.255099961534 }, { "content_chunks": 65, "decode_tok_s": 120.27239461313785, "e2e_ms": 5231.222199974582, "effective_prefill_tok_s": 3733.2071834352178, "generation_ms": 4248.689000029117, "mean_inter_token_latency_ms": 8.31445988264015, "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": 982.5331999454647, "tpot_ms": 8.31445988264015, "ttft_ms": 982.5334999477491, "visible_content_truncated": false, "visible_generation_ms": 4248.6887000268325 }, { "content_chunks": 65, "decode_tok_s": 120.33234711113103, "e2e_ms": 5474.229699932039, "effective_prefill_tok_s": 2988.6185683968024, "generation_ms": 4246.572199976072, "mean_inter_token_latency_ms": 8.310317416782919, "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": 1227.657499955967, "tpot_ms": 8.310317416782919, "ttft_ms": 1227.6577999582514, "visible_content_truncated": false, "visible_generation_ms": 4246.571899973787 }, { "content_chunks": 56, "decode_tok_s": 105.91091201608836, "e2e_ms": 6099.015600048006, "effective_prefill_tok_s": 1895.9881272996822, "generation_ms": 4163.876899983734, "mean_inter_token_latency_ms": 9.44189773238942, "output_token_source": "usage", "output_tokens": 442, "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": 1935.1387000642717, "tpot_ms": 9.44189773238942, "ttft_ms": 1935.1388000650331, "visible_content_truncated": false, "visible_generation_ms": 4163.876799982972 }, { "content_chunks": 65, "decode_tok_s": 110.29743548824226, "e2e_ms": 5463.976200087927, "effective_prefill_tok_s": 4416.105959436682, "generation_ms": 4632.927300059237, "mean_inter_token_latency_ms": 9.06639393357972, "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": 831.0489000286907, "tpot_ms": 9.06639393357972, "ttft_ms": 831.049200030975, "visible_content_truncated": false, "visible_generation_ms": 4632.927000056952 }, { "content_chunks": 65, "decode_tok_s": 101.74039184851588, "e2e_ms": 6288.76260004472, "effective_prefill_tok_s": 2899.282587491651, "generation_ms": 5022.587300045416, "mean_inter_token_latency_ms": 9.828937964863828, "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": 1266.1752999993041, "tpot_ms": 9.828937964863828, "ttft_ms": 1266.1756000015885, "visible_content_truncated": false, "visible_generation_ms": 5022.587000043131 }, { "content_chunks": 35, "decode_tok_s": 122.07039652496725, "e2e_ms": 3137.4676000559703, "effective_prefill_tok_s": 3831.397236148224, "generation_ms": 2179.070500074886, "mean_inter_token_latency_ms": 8.19199436118378, "output_token_source": "usage", "output_tokens": 267, "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": 958.3970999810845, "tpot_ms": 8.19199436118378, "ttft_ms": 958.3973999833688, "visible_content_truncated": false, "visible_generation_ms": 2179.0702000726014 }, { "content_chunks": 46, "decode_tok_s": 123.53934565786722, "e2e_ms": 3851.547300000675, "effective_prefill_tok_s": 3850.32753491021, "generation_ms": 2897.862200043164, "mean_inter_token_latency_ms": 8.09458715095856, "output_token_source": "usage", "output_tokens": 359, "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": 953.6850999575108, "tpot_ms": 8.09458715095856, "ttft_ms": 953.6853999597952, "visible_content_truncated": false, "visible_generation_ms": 2897.86190004088 }, { "content_chunks": 65, "decode_tok_s": 120.13609844938779, "e2e_ms": 5264.190800022334, "effective_prefill_tok_s": 3633.1916992650918, "generation_ms": 4253.509199945256, "mean_inter_token_latency_ms": 8.323892759188368, "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": 1010.6816000770777, "tpot_ms": 8.323892759188368, "ttft_ms": 1010.6819000793621, "visible_content_truncated": false, "visible_generation_ms": 4253.508899942972 }, { "content_chunks": 65, "decode_tok_s": 119.94066903659669, "e2e_ms": 5482.287400052883, "effective_prefill_tok_s": 3005.284783296499, "generation_ms": 4260.439799982123, "mean_inter_token_latency_ms": 8.337455577264429, "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": 1221.8476000707597, "tpot_ms": 8.337455577264429, "ttft_ms": 1221.847900073044, "visible_content_truncated": false, "visible_generation_ms": 4260.439499979839 }, { "content_chunks": 61, "decode_tok_s": 115.82442522830411, "e2e_ms": 5366.971499985084, "effective_prefill_tok_s": 3003.0228958978555, "generation_ms": 4144.203600008041, "mean_inter_token_latency_ms": 8.63375750001675, "output_token_source": "usage", "output_tokens": 481, "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": 1222.7678999770433, "tpot_ms": 8.63375750001675, "ttft_ms": 1222.768299980089, "visible_content_truncated": false, "visible_generation_ms": 4144.203200004995 }, { "content_chunks": 65, "decode_tok_s": 119.72690550783955, "e2e_ms": 5339.8237999062985, "effective_prefill_tok_s": 3426.084877964784, "generation_ms": 4268.0464999284595, "mean_inter_token_latency_ms": 8.352341487139842, "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": 1071.777299977839, "tpot_ms": 8.352341487139842, "ttft_ms": 1071.777499979362, "visible_content_truncated": false, "visible_generation_ms": 4268.046299926937 }, { "content_chunks": 65, "decode_tok_s": 126.45302193720545, "e2e_ms": 5316.200400004163, "effective_prefill_tok_s": 2881.1754315921808, "generation_ms": 4041.026399936527, "mean_inter_token_latency_ms": 7.908075146646824, "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": 1275.1740000676364, "tpot_ms": 7.908075146646824, "ttft_ms": 1275.1747999573126, "visible_content_truncated": false, "visible_generation_ms": 4041.0256000468507 }, { "content_chunks": 65, "decode_tok_s": 131.59744881300256, "e2e_ms": 4719.805600005202, "effective_prefill_tok_s": 4390.789333265571, "generation_ms": 3883.053999976255, "mean_inter_token_latency_ms": 7.598931506802847, "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": 836.751600028947, "tpot_ms": 7.598931506802847, "ttft_ms": 836.7520000319928, "visible_content_truncated": false, "visible_generation_ms": 3883.053599973209 }, { "content_chunks": 65, "decode_tok_s": 119.73463150233461, "e2e_ms": 5797.852000105195, "effective_prefill_tok_s": 2401.8337852929117, "generation_ms": 4267.771100043319, "mean_inter_token_latency_ms": 8.351802544116085, "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": 1530.0809000618756, "tpot_ms": 8.351802544116085, "ttft_ms": 1530.08120006416, "visible_content_truncated": false, "visible_generation_ms": 4267.770800041035 }, { "content_chunks": 65, "decode_tok_s": 119.70828753657543, "e2e_ms": 5532.056399970315, "effective_prefill_tok_s": 2909.7331285700297, "generation_ms": 4268.710299977101, "mean_inter_token_latency_ms": 8.353640508761451, "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": 1263.346099993214, "tpot_ms": 8.353640508761451, "ttft_ms": 1263.3463999954984, "visible_content_truncated": false, "visible_generation_ms": 4268.709999974817 }, { "content_chunks": 65, "decode_tok_s": 119.65952066074097, "e2e_ms": 5502.693899907172, "effective_prefill_tok_s": 2983.9871799209736, "generation_ms": 4270.449999952689, "mean_inter_token_latency_ms": 8.35704500969215, "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": 1232.243899954483, "tpot_ms": 8.35704500969215, "ttft_ms": 1232.2441999567673, "visible_content_truncated": false, "visible_generation_ms": 4270.4496999504045 }, { "content_chunks": 65, "decode_tok_s": 119.37891751448764, "e2e_ms": 5389.1504999483, "effective_prefill_tok_s": 3316.6083789306226, "generation_ms": 4280.487800016999, "mean_inter_token_latency_ms": 8.376688454045008, "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": 1108.6626999313012, "tpot_ms": 8.376688454045008, "ttft_ms": 1108.6629999335855, "visible_content_truncated": false, "visible_generation_ms": 4280.487500014715 }, { "content_chunks": 65, "decode_tok_s": 119.39748617249896, "e2e_ms": 5480.338099994697, "effective_prefill_tok_s": 3063.682616370778, "generation_ms": 4279.82209995389, "mean_inter_token_latency_ms": 8.37538571419548, "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": 1200.5160000408068, "tpot_ms": 8.37538571419548, "ttft_ms": 1200.5164000438526, "visible_content_truncated": false, "visible_generation_ms": 4279.821699950844 }, { "content_chunks": 65, "decode_tok_s": 125.96594743385586, "e2e_ms": 5338.339500012808, "effective_prefill_tok_s": 2870.4342619263325, "generation_ms": 4056.6519000567496, "mean_inter_token_latency_ms": 7.9386534247685905, "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": 1281.687599956058, "tpot_ms": 7.9386534247685905, "ttft_ms": 1281.6879999591038, "visible_content_truncated": false, "visible_generation_ms": 4056.651500053704 }, { "content_chunks": 65, "decode_tok_s": 131.04083840437664, "e2e_ms": 4735.012999968603, "effective_prefill_tok_s": 4404.731112041838, "generation_ms": 3899.547699955292, "mean_inter_token_latency_ms": 7.631208806174739, "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": 835.4653000133112, "tpot_ms": 7.631208806174739, "ttft_ms": 835.4656000155956, "visible_content_truncated": false, "visible_generation_ms": 3899.5473999530077 }, { "content_chunks": 65, "decode_tok_s": 119.16782241153612, "e2e_ms": 5845.716000068933, "effective_prefill_tok_s": 2361.897830723177, "generation_ms": 4288.070300011896, "mean_inter_token_latency_ms": 8.39152700589412, "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": 1557.6457000570372, "tpot_ms": 8.39152700589412, "ttft_ms": 1557.6460000593215, "visible_content_truncated": false, "visible_generation_ms": 4288.070000009611 }, { "content_chunks": 65, "decode_tok_s": 119.00797467603672, "e2e_ms": 5558.702899958007, "effective_prefill_tok_s": 2909.382997394421, "generation_ms": 4293.829899979755, "mean_inter_token_latency_ms": 8.402798238707936, "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": 1264.8729999782518, "tpot_ms": 8.402798238707936, "ttft_ms": 1264.873499982059, "visible_content_truncated": false, "visible_generation_ms": 4293.829399975948 }, { "content_chunks": 65, "decode_tok_s": 119.03762159609877, "e2e_ms": 5530.844899942167, "effective_prefill_tok_s": 2973.1414111428835, "generation_ms": 4292.760499985889, "mean_inter_token_latency_ms": 8.40070547942444, "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": 1238.0843999562785, "tpot_ms": 8.40070547942444, "ttft_ms": 1238.0846999585629, "visible_content_truncated": false, "visible_generation_ms": 4292.760199983604 }, { "content_chunks": 65, "decode_tok_s": 119.03306852908058, "e2e_ms": 5412.32330002822, "effective_prefill_tok_s": 3289.266218272204, "generation_ms": 4292.92469995562, "mean_inter_token_latency_ms": 8.401026810089276, "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": 1119.3986000726, "tpot_ms": 8.401026810089276, "ttft_ms": 1119.39930007793, "visible_content_truncated": false, "visible_generation_ms": 4292.92399995029 }, { "content_chunks": 65, "decode_tok_s": 119.04090767369537, "e2e_ms": 5500.750000006519, "effective_prefill_tok_s": 3048.5685054858754, "generation_ms": 4292.642000014894, "mean_inter_token_latency_ms": 8.400473581242453, "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": 1208.1079999916255, "tpot_ms": 8.400473581242453, "ttft_ms": 1208.1086999969557, "visible_content_truncated": false, "visible_generation_ms": 4292.641300009564 }, { "content_chunks": 65, "decode_tok_s": 125.80939634936239, "e2e_ms": 5348.831100040115, "effective_prefill_tok_s": 2862.1788624868873, "generation_ms": 4061.6998000768945, "mean_inter_token_latency_ms": 7.948531898389226, "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": 1287.131299963221, "tpot_ms": 7.948531898389226, "ttft_ms": 1287.1316999662668, "visible_content_truncated": false, "visible_generation_ms": 4061.6994000738487 }, { "content_chunks": 65, "decode_tok_s": 145.87942440269418, "e2e_ms": 5067.970099975355, "effective_prefill_tok_s": 2354.516441820251, "generation_ms": 3502.892900025472, "mean_inter_token_latency_ms": 6.8549763209891825, "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": 1565.077199949883, "tpot_ms": 6.8549763209891825, "ttft_ms": 1565.0775999529287, "visible_content_truncated": false, "visible_generation_ms": 3502.8925000224262 }, { "content_chunks": 65, "decode_tok_s": 130.9183383403278, "e2e_ms": 4741.361299995333, "effective_prefill_tok_s": 4397.703172648349, "generation_ms": 3903.196500032209, "mean_inter_token_latency_ms": 7.638349315131524, "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": 838.1647999631241, "tpot_ms": 7.638349315131524, "ttft_ms": 838.1650999654084, "visible_content_truncated": false, "visible_generation_ms": 3903.1962000299245 }, { "content_chunks": 65, "decode_tok_s": 165.3205443982829, "e2e_ms": 4363.504300010391, "effective_prefill_tok_s": 2896.5704325151096, "generation_ms": 3090.9649000968784, "mean_inter_token_latency_ms": 6.04885499040485, "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": 1272.5393999135122, "tpot_ms": 6.04885499040485, "ttft_ms": 1272.5396999157965, "visible_content_truncated": false, "visible_generation_ms": 3090.964600094594 }, { "content_chunks": 65, "decode_tok_s": 191.34227505404303, "e2e_ms": 3909.5681000035256, "effective_prefill_tok_s": 2975.880116070229, "generation_ms": 2670.6068998901173, "mean_inter_token_latency_ms": 5.2262365946969025, "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": 1238.9612001134083, "tpot_ms": 5.2262365946969025, "ttft_ms": 1238.9617000008002, "visible_content_truncated": false, "visible_generation_ms": 2670.6064000027254 }, { "content_chunks": 65, "decode_tok_s": 226.87510050545563, "e2e_ms": 3370.21189997904, "effective_prefill_tok_s": 3300.0221046065562, "generation_ms": 2252.340600010939, "mean_inter_token_latency_ms": 4.407711546009665, "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": 1117.871299968101, "tpot_ms": 4.407711546009665, "ttft_ms": 1117.8716999711469, "visible_content_truncated": false, "visible_generation_ms": 2252.340200007893 }, { "content_chunks": 65, "decode_tok_s": 281.94148771362916, "e2e_ms": 3018.458499922417, "effective_prefill_tok_s": 3058.807121686135, "generation_ms": 1812.4327999539673, "mean_inter_token_latency_ms": 3.54683522495884, "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": 1206.0256999684498, "tpot_ms": 3.54683522495884, "ttft_ms": 1206.0258999699727, "visible_content_truncated": false, "visible_generation_ms": 1812.4325999524444 } ], "requests": 100, "response_words": 256, "run_id": "benchmark-20260905T042829Z", "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": "2026-09-05T04:28:29Z", "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": 66212.15000003576 }