{ "base_url": "http://127.0.0.1:39277/v1", "completed": 100, "concurrency": 8, "context_distribution": { "4096": 100 }, "context_seed": 0, "context_tokens": 4096, "engine": "sglang", "failed": 0, "failures": [], "finished_at": "2026-09-05T04:11:56Z", "gpu": "2x 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": 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": 4459, "content_chunks_s": null, "decode_tok_s_max": 356.2181041683724, "decode_tok_s_mean": 167.0476547301582, "decode_tok_s_mean_ci95_high": 178.906085651566, "decode_tok_s_mean_ci95_low": 155.1892238087504, "decode_tok_s_min": 97.82246990566618, "decode_tok_s_p25": 113.7398393705312, "decode_tok_s_p50": 150.16954758710438, "decode_tok_s_p75": 217.83059821669957, "decode_tok_s_p90": 266.504559569707, "decode_tok_s_p95": 274.74039989645, "decode_tok_s_p99": 276.27865520180313, "decode_tok_s_samples": 100, "decode_tok_s_stddev": 60.502198578611164, "e2e_max_ms": 12081.020699930377, "e2e_mean_ci95_high_ms": 7206.066465929292, "e2e_mean_ci95_low_ms": 6404.242862079152, "e2e_mean_ms": 6805.154664004222, "e2e_min_ms": 2725.225300062448, "e2e_p25_ms": 6007.502100081183, "e2e_p50_ms": 6046.960200066678, "e2e_p75_ms": 6146.993099944666, "e2e_p90_ms": 11849.902300047688, "e2e_p95_ms": 11971.510399947874, "e2e_p99_ms": 12072.478399961255, "e2e_samples": 100, "e2e_stddev_ms": 2045.4683771687235, "effective_prefill_tok_s_max": 3408.4360915958264, "effective_prefill_tok_s_mean": 1367.5248779549981, "effective_prefill_tok_s_mean_ci95_high": 1498.474400362279, "effective_prefill_tok_s_mean_ci95_low": 1236.5753555477172, "effective_prefill_tok_s_min": 519.9772771958305, "effective_prefill_tok_s_p25": 921.6257174458036, "effective_prefill_tok_s_p50": 1149.4599729485933, "effective_prefill_tok_s_p75": 1772.12536203138, "effective_prefill_tok_s_p90": 2428.614903751655, "effective_prefill_tok_s_p95": 2804.987085311303, "effective_prefill_tok_s_p99": 2858.114857525705, "effective_prefill_tok_s_samples": 100, "effective_prefill_tok_s_stddev": 668.1098082004125, "generation_max_ms": 5223.748699994758, "generation_mean_ci95_high_ms": 3664.0776045653492, "generation_mean_ci95_low_ms": 3223.3540234518714, "generation_mean_ms": 3443.7158140086103, "generation_min_ms": 1434.5144000835717, "generation_p25_ms": 2342.2501001041383, "generation_p50_ms": 3351.8723000306636, "generation_p75_ms": 4475.751699996181, "generation_p90_ms": 5138.722800067626, "generation_p95_ms": 5176.567599992268, "generation_p99_ms": 5218.969199922867, "generation_samples": 100, "generation_stddev_ms": 1124.2948497792795, "mean_inter_token_latency_ms_max": 10.222600195684457, "mean_inter_token_latency_ms_mean": 6.739169890427808, "mean_inter_token_latency_ms_mean_ci95_high": 7.170406271165066, "mean_inter_token_latency_ms_mean_ci95_low": 6.3079335096905504, "mean_inter_token_latency_ms_min": 2.807268884703663, "mean_inter_token_latency_ms_p25": 4.583659687092247, "mean_inter_token_latency_ms_p50": 6.559436986361377, "mean_inter_token_latency_ms_p75": 8.758809589033621, "mean_inter_token_latency_ms_p90": 10.056209002089288, "mean_inter_token_latency_ms_p95": 10.130269275914419, "mean_inter_token_latency_ms_p99": 10.213246966580954, "mean_inter_token_latency_ms_samples": 100, "mean_inter_token_latency_ms_stddev": 2.200185616006418, "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": 587.9371362276845, "time_to_first_output_max_ms": 6965.6889999751, "time_to_first_output_mean_ci95_high_ms": 3684.596970077914, "time_to_first_output_mean_ci95_low_ms": 3038.2807299133083, "time_to_first_output_mean_ms": 3361.4388499956112, "time_to_first_output_min_ms": 1061.1904999241233, "time_to_first_output_p25_ms": 1989.0668999869376, "time_to_first_output_p50_ms": 3184.3951999908313, "time_to_first_output_p75_ms": 3956.052799941972, "time_to_first_output_p90_ms": 6690.39210001938, "time_to_first_output_p95_ms": 6826.201199903153, "time_to_first_output_p99_ms": 6890.053999959491, "time_to_first_output_samples": 100, "time_to_first_output_stddev_ms": 1648.76591878726, "tpot_max_ms": 10.222600195684457, "tpot_mean_ci95_high_ms": 7.170406271165066, "tpot_mean_ci95_low_ms": 6.3079335096905504, "tpot_mean_ms": 6.739169890427808, "tpot_min_ms": 2.807268884703663, "tpot_p25_ms": 4.583659687092247, "tpot_p50_ms": 6.559436986361377, "tpot_p75_ms": 8.758809589033621, "tpot_p90_ms": 10.056209002089288, "tpot_p95_ms": 10.130269275914419, "tpot_p99_ms": 10.213246966580954, "tpot_samples": 100, "tpot_stddev_ms": 2.200185616006418, "ttft_max_ms": 6965.689499978907, "ttft_mean_ci95_high_ms": 3684.597465979297, "ttft_mean_ci95_low_ms": 3038.2812260078376, "ttft_mean_ms": 3361.4393459935673, "ttft_min_ms": 1061.191099928692, "ttft_p25_ms": 1989.0672999899834, "ttft_p50_ms": 3184.395599993877, "ttft_p75_ms": 3956.053299945779, "ttft_p90_ms": 6690.392700023949, "ttft_p95_ms": 6826.20169990696, "ttft_p99_ms": 6890.054399962537, "ttft_samples": 100, "ttft_stddev_ms": 1648.765918294539 }, "model": "qwen38-27b-inferact-nvfp4-pro6000-tp2-dflash2-k12-chunk1k", "prompt_cache_mode": "unique", "request_canaries": true, "request_timings": [ { "content_chunks": 44, "decode_tok_s": 130.28592431468067, "e2e_ms": 6104.576100013219, "effective_prefill_tok_s": 1655.49160200113, "generation_ms": 3922.142800060101, "mean_inter_token_latency_ms": 7.675426223209591, "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": 2182.433299953118, "tpot_ms": 7.675426223209591, "ttft_ms": 2182.4335999554023, "visible_content_truncated": false, "visible_generation_ms": 3922.1425000578165 }, { "content_chunks": 44, "decode_tok_s": 116.07474113917155, "e2e_ms": 6104.689100058749, "effective_prefill_tok_s": 2123.5311214335065, "generation_ms": 4402.335899998434, "mean_inter_token_latency_ms": 8.615138747550752, "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": 1702.3532000603154, "tpot_ms": 8.615138747550752, "ttft_ms": 1702.3539000656456, "visible_content_truncated": false, "visible_generation_ms": 4402.335199993104 }, { "content_chunks": 44, "decode_tok_s": 101.34933047563473, "e2e_ms": 6103.157699923031, "effective_prefill_tok_s": 3408.4360915958264, "generation_ms": 5041.967199998908, "mean_inter_token_latency_ms": 9.866863405085924, "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": 1061.1904999241233, "tpot_ms": 9.866863405085924, "ttft_ms": 1061.191099928692, "visible_content_truncated": false, "visible_generation_ms": 5041.966599994339 }, { "content_chunks": 44, "decode_tok_s": 182.34496119585754, "e2e_ms": 6104.552400065586, "effective_prefill_tok_s": 1095.3397728964421, "generation_ms": 2802.3807000136003, "mean_inter_token_latency_ms": 5.484110958930724, "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": 3302.1717000519857, "tpot_ms": 5.484110958930724, "ttft_ms": 3302.1721000550315, "visible_content_truncated": false, "visible_generation_ms": 2802.3803000105545 }, { "content_chunks": 45, "decode_tok_s": 153.64174533032875, "e2e_ms": 6147.274099988863, "effective_prefill_tok_s": 1282.7169469452112, "generation_ms": 3325.9190000826493, "mean_inter_token_latency_ms": 6.5086477496725035, "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": 2821.3550999062136, "tpot_ms": 6.5086477496725035, "ttft_ms": 2821.3556999107823, "visible_content_truncated": false, "visible_generation_ms": 3325.9184000780806 }, { "content_chunks": 45, "decode_tok_s": 231.25727656576686, "e2e_ms": 6146.993099944666, "effective_prefill_tok_s": 919.4040971489878, "generation_ms": 2209.6602000528947, "mean_inter_token_latency_ms": 4.324188258420538, "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": 3937.3328998917714, "tpot_ms": 4.324188258420538, "ttft_ms": 3937.3333998955786, "visible_content_truncated": false, "visible_generation_ms": 2209.6597000490874 }, { "content_chunks": 44, "decode_tok_s": 275.44782101289167, "e2e_ms": 6102.183100068942, "effective_prefill_tok_s": 852.8328188850594, "generation_ms": 1855.1608000416309, "mean_inter_token_latency_ms": 3.6304516634865576, "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": 4247.022300027311, "tpot_ms": 3.6304516634865576, "ttft_ms": 4247.022700030357, "visible_content_truncated": false, "visible_generation_ms": 1855.160400038585 }, { "content_chunks": 44, "decode_tok_s": 100.06286924645558, "e2e_ms": 12072.478399961255, "effective_prefill_tok_s": 519.9772771958305, "generation_ms": 5106.789399986155, "mean_inter_token_latency_ms": 9.99371702541322, "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": 6965.6889999751, "tpot_ms": 9.99371702541322, "ttft_ms": 6965.689499978907, "visible_content_truncated": false, "visible_generation_ms": 5106.788899982348 }, { "content_chunks": 45, "decode_tok_s": 126.65428214365478, "e2e_ms": 6013.716599904001, "effective_prefill_tok_s": 1830.1140774765797, "generation_ms": 4034.6049999352545, "mean_inter_token_latency_ms": 7.895508806135528, "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": 1979.1115999687463, "tpot_ms": 7.895508806135528, "ttft_ms": 1979.1122999740764, "visible_content_truncated": false, "visible_generation_ms": 4034.6042999299243 }, { "content_chunks": 44, "decode_tok_s": 177.90373005970574, "e2e_ms": 5969.790200004354, "effective_prefill_tok_s": 1169.9945974908364, "generation_ms": 2872.3400000017136, "mean_inter_token_latency_ms": 5.621017612527815, "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": 3097.4502000026405, "tpot_ms": 5.621017612527815, "ttft_ms": 3097.4506000056863, "visible_content_truncated": false, "visible_generation_ms": 2872.339599998668 }, { "content_chunks": 45, "decode_tok_s": 113.31988042249996, "e2e_ms": 6013.076800038107, "effective_prefill_tok_s": 2411.35638460637, "generation_ms": 4509.358800016344, "mean_inter_token_latency_ms": 8.824576908055468, "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": 1503.7180000217631, "tpot_ms": 8.824576908055468, "ttft_ms": 1503.718400024809, "visible_content_truncated": false, "visible_generation_ms": 4509.358400013298 }, { "content_chunks": 44, "decode_tok_s": 209.1919345388485, "e2e_ms": 5969.1773999948055, "effective_prefill_tok_s": 1028.514668864303, "generation_ms": 2442.732800031081, "mean_inter_token_latency_ms": 4.7802990215872425, "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": 3526.4445999637246, "tpot_ms": 4.7802990215872425, "ttft_ms": 3526.4451999682933, "visible_content_truncated": false, "visible_generation_ms": 2442.732200026512 }, { "content_chunks": 44, "decode_tok_s": 152.74424769227323, "e2e_ms": 5969.217300065793, "effective_prefill_tok_s": 1382.3695024896472, "generation_ms": 3345.4614999936894, "mean_inter_token_latency_ms": 6.546891389420136, "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": 2623.7558000721037, "tpot_ms": 6.546891389420136, "ttft_ms": 2623.7566999625415, "visible_content_truncated": false, "visible_generation_ms": 3345.4606001032516 }, { "content_chunks": 44, "decode_tok_s": 274.3852655991864, "e2e_ms": 5968.82499998901, "effective_prefill_tok_s": 883.4816951707882, "generation_ms": 1862.344899913296, "mean_inter_token_latency_ms": 3.644510567345002, "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": 4106.480100075714, "tpot_ms": 3.644510567345002, "ttft_ms": 4106.480400077999, "visible_content_truncated": false, "visible_generation_ms": 1862.3445999110118 }, { "content_chunks": 44, "decode_tok_s": 99.44105177132248, "e2e_ms": 11928.667800035328, "effective_prefill_tok_s": 534.4667740338489, "generation_ms": 5138.722800067626, "mean_inter_token_latency_ms": 10.056209002089288, "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": 6789.944999967702, "tpot_ms": 10.056209002089288, "ttft_ms": 6789.9453999707475, "visible_content_truncated": false, "visible_generation_ms": 5138.72240006458 }, { "content_chunks": 45, "decode_tok_s": 149.1343524207361, "e2e_ms": 6044.651000062004, "effective_prefill_tok_s": 1386.443198000893, "generation_ms": 3426.440600072965, "mean_inter_token_latency_ms": 6.705363209536135, "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": 2618.2103999890387, "tpot_ms": 6.705363209536135, "ttft_ms": 2618.210899992846, "visible_content_truncated": false, "visible_generation_ms": 3426.440100069158 }, { "content_chunks": 44, "decode_tok_s": 113.48778569159795, "e2e_ms": 6003.869600011967, "effective_prefill_tok_s": 2417.4277557385335, "generation_ms": 4502.687200088985, "mean_inter_token_latency_ms": 8.811520939508778, "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": 1501.1823999229819, "tpot_ms": 8.811520939508778, "ttft_ms": 1501.182899926789, "visible_content_truncated": false, "visible_generation_ms": 4502.686700085178 }, { "content_chunks": 45, "decode_tok_s": 125.84815747159897, "e2e_ms": 6044.0047000302, "effective_prefill_tok_s": 1830.0467358256194, "generation_ms": 4060.4488000972196, "mean_inter_token_latency_ms": 7.946083757528806, "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": 1983.5558999329805, "tpot_ms": 7.946083757528806, "ttft_ms": 1983.5562999360263, "visible_content_truncated": false, "visible_generation_ms": 4060.448400094174 }, { "content_chunks": 44, "decode_tok_s": 165.9825435768738, "e2e_ms": 6002.933299983852, "effective_prefill_tok_s": 1241.6662428995896, "generation_ms": 3078.6369999404997, "mean_inter_token_latency_ms": 6.0247299411751465, "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": 2924.296300043352, "tpot_ms": 6.0247299411751465, "ttft_ms": 2924.2967000463977, "visible_content_truncated": false, "visible_generation_ms": 3078.636599937454 }, { "content_chunks": 45, "decode_tok_s": 205.91814386653527, "e2e_ms": 6185.4148000711575, "effective_prefill_tok_s": 980.6022723028663, "generation_ms": 2481.568600051105, "mean_inter_token_latency_ms": 4.856298630236996, "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": 3703.8462000200525, "tpot_ms": 4.856298630236996, "ttft_ms": 3703.8469000253826, "visible_content_truncated": false, "visible_generation_ms": 2481.567900045775 }, { "content_chunks": 45, "decode_tok_s": 99.10498501293785, "e2e_ms": 11982.349499943666, "effective_prefill_tok_s": 532.2140226472585, "generation_ms": 5156.148300040513, "mean_inter_token_latency_ms": 10.090309784815094, "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": 6826.201199903153, "tpot_ms": 10.090309784815094, "ttft_ms": 6826.20169990696, "visible_content_truncated": false, "visible_generation_ms": 5156.147800036706 }, { "content_chunks": 45, "decode_tok_s": 249.48751310709358, "e2e_ms": 6184.757299954072, "effective_prefill_tok_s": 878.5080428928173, "generation_ms": 2048.198699951172, "mean_inter_token_latency_ms": 4.0082166339553265, "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": 4136.5586000029, "tpot_ms": 4.0082166339553265, "ttft_ms": 4136.559000005946, "visible_content_truncated": false, "visible_generation_ms": 2048.198299948126 }, { "content_chunks": 44, "decode_tok_s": 114.24335385947316, "e2e_ms": 5978.43469993677, "effective_prefill_tok_s": 2414.4372588660663, "generation_ms": 4472.907899995334, "mean_inter_token_latency_ms": 8.753244422691457, "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": 1505.5267999414355, "tpot_ms": 8.753244422691457, "ttft_ms": 1505.5273999460042, "visible_content_truncated": false, "visible_generation_ms": 4472.907299990766 }, { "content_chunks": 45, "decode_tok_s": 141.4447241643425, "e2e_ms": 6021.68760006316, "effective_prefill_tok_s": 1509.3594607438833, "generation_ms": 3612.718699965626, "mean_inter_token_latency_ms": 7.0698996085433, "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": 2408.968900097534, "tpot_ms": 7.0698996085433, "ttft_ms": 2408.96930010058, "visible_content_truncated": false, "visible_generation_ms": 3612.71829996258 }, { "content_chunks": 45, "decode_tok_s": 126.4413290833306, "e2e_ms": 6020.911899977364, "effective_prefill_tok_s": 1836.8165322478565, "generation_ms": 4041.4000999880955, "mean_inter_token_latency_ms": 7.908806457902339, "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": 1979.5117999892682, "tpot_ms": 7.908806457902339, "ttft_ms": 1979.5122999930754, "visible_content_truncated": false, "visible_generation_ms": 4041.399599984288 }, { "content_chunks": 45, "decode_tok_s": 171.24368187150333, "e2e_ms": 6168.446899973787, "effective_prefill_tok_s": 1142.132107224151, "generation_ms": 2984.051699982956, "mean_inter_token_latency_ms": 5.8396315068159605, "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": 3184.3951999908313, "tpot_ms": 5.8396315068159605, "ttft_ms": 3184.395599993877, "visible_content_truncated": false, "visible_generation_ms": 2984.0512999799103 }, { "content_chunks": 45, "decode_tok_s": 217.83059821669957, "e2e_ms": 6168.565100058913, "effective_prefill_tok_s": 951.9435907218655, "generation_ms": 2345.859599998221, "mean_inter_token_latency_ms": 4.5907232876677515, "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": 3822.7055000606924, "tpot_ms": 4.5907232876677515, "ttft_ms": 3822.706100065261, "visible_content_truncated": false, "visible_generation_ms": 2345.858999993652 }, { "content_chunks": 45, "decode_tok_s": 250.4596743809718, "e2e_ms": 5984.957900014706, "effective_prefill_tok_s": 922.5014375435408, "generation_ms": 2040.2485999511555, "mean_inter_token_latency_ms": 3.992658708319287, "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": 3944.7093000635505, "tpot_ms": 3.992658708319287, "ttft_ms": 3944.7097000665963, "visible_content_truncated": false, "visible_generation_ms": 2040.2481999481097 }, { "content_chunks": 45, "decode_tok_s": 98.48322344911315, "e2e_ms": 11813.429199974053, "effective_prefill_tok_s": 549.4565045022916, "generation_ms": 5188.701000064611, "mean_inter_token_latency_ms": 10.154013698756579, "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": 6624.728199909441, "tpot_ms": 10.154013698756579, "ttft_ms": 6624.728599912487, "visible_content_truncated": false, "visible_generation_ms": 5188.700600061566 }, { "content_chunks": 45, "decode_tok_s": 107.29511226700109, "e2e_ms": 6060.967199970037, "effective_prefill_tok_s": 2804.987085311303, "generation_ms": 4762.565499986522, "mean_inter_token_latency_ms": 9.320089041069515, "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": 1298.4016999835148, "tpot_ms": 9.320089041069515, "ttft_ms": 1298.402199987322, "visible_content_truncated": false, "visible_generation_ms": 4762.564999982715 }, { "content_chunks": 45, "decode_tok_s": 152.49586186918853, "e2e_ms": 6062.306700041518, "effective_prefill_tok_s": 1343.2194580498985, "generation_ms": 3350.9106000419706, "mean_inter_token_latency_ms": 6.557554990297398, "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": 2711.396099999547, "tpot_ms": 6.557554990297398, "ttft_ms": 2711.396500002593, "visible_content_truncated": false, "visible_generation_ms": 3350.910200038925 }, { "content_chunks": 45, "decode_tok_s": 177.85155117258108, "e2e_ms": 6060.900999931619, "effective_prefill_tok_s": 1142.823693077265, "generation_ms": 2873.1827000156045, "mean_inter_token_latency_ms": 5.622666731928776, "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": 3187.718299916014, "tpot_ms": 5.622666731928776, "ttft_ms": 3187.7185999182984, "visible_content_truncated": false, "visible_generation_ms": 2873.18240001332 }, { "content_chunks": 44, "decode_tok_s": 129.3634893558753, "e2e_ms": 6016.813699970953, "effective_prefill_tok_s": 1763.194125838337, "generation_ms": 3950.1098999753594, "mean_inter_token_latency_ms": 7.730156360030057, "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": 2066.7037999955937, "tpot_ms": 7.730156360030057, "ttft_ms": 2066.7041999986395, "visible_content_truncated": false, "visible_generation_ms": 3950.1094999723136 }, { "content_chunks": 45, "decode_tok_s": 99.04040891269028, "e2e_ms": 11849.902300047688, "effective_prefill_tok_s": 544.8111180194419, "generation_ms": 5159.510200028308, "mean_inter_token_latency_ms": 10.09688884545657, "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": 6690.39210001938, "tpot_ms": 10.09688884545657, "ttft_ms": 6690.392700023949, "visible_content_truncated": false, "visible_generation_ms": 5159.509600023739 }, { "content_chunks": 45, "decode_tok_s": 228.5299775617875, "e2e_ms": 5873.873500037007, "effective_prefill_tok_s": 1002.2422627903192, "generation_ms": 2236.030500032939, "mean_inter_token_latency_ms": 4.375793542138824, "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": 3637.843000004068, "tpot_ms": 4.375793542138824, "ttft_ms": 3637.8435000078753, "visible_content_truncated": false, "visible_generation_ms": 2236.0300000291318 }, { "content_chunks": 45, "decode_tok_s": 266.504559569707, "e2e_ms": 5873.468399979174, "effective_prefill_tok_s": 921.6257174458036, "generation_ms": 1917.4156000372022, "mean_inter_token_latency_ms": 3.7522810176853274, "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": 3956.052799941972, "tpot_ms": 3.7522810176853274, "ttft_ms": 3956.053299945779, "visible_content_truncated": false, "visible_generation_ms": 1917.415100033395 }, { "content_chunks": 44, "decode_tok_s": 114.17076599679022, "e2e_ms": 5977.43069997523, "effective_prefill_tok_s": 2428.614903751655, "generation_ms": 4475.751699996181, "mean_inter_token_latency_ms": 8.758809589033621, "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": 1501.678999979049, "tpot_ms": 8.758809589033621, "ttft_ms": 1501.6795999836177, "visible_content_truncated": false, "visible_generation_ms": 4475.751099991612 }, { "content_chunks": 45, "decode_tok_s": 126.45934329914309, "e2e_ms": 6020.092899911106, "effective_prefill_tok_s": 1843.6104045516197, "generation_ms": 4040.8243999117985, "mean_inter_token_latency_ms": 7.907679843271621, "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": 1979.268499999307, "tpot_ms": 7.907679843271621, "ttft_ms": 1979.268900002353, "visible_content_truncated": false, "visible_generation_ms": 4040.8239999087527 }, { "content_chunks": 44, "decode_tok_s": 143.30804839375443, "e2e_ms": 5977.0777999656275, "effective_prefill_tok_s": 1513.271189237968, "generation_ms": 3565.745299914852, "mean_inter_token_latency_ms": 6.977975146604407, "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": 2411.3325000507757, "tpot_ms": 6.977975146604407, "ttft_ms": 2411.3331000553444, "visible_content_truncated": false, "visible_generation_ms": 3565.744699910283 }, { "content_chunks": 45, "decode_tok_s": 171.52939062287382, "e2e_ms": 6162.544499966316, "effective_prefill_tok_s": 1146.550084199248, "generation_ms": 2979.081299970858, "mean_inter_token_latency_ms": 5.82990469661616, "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": 3183.463199995458, "tpot_ms": 5.82990469661616, "ttft_ms": 3183.463599998504, "visible_content_truncated": false, "visible_generation_ms": 2979.080899967812 }, { "content_chunks": 44, "decode_tok_s": 275.78786547215486, "e2e_ms": 5976.0360000655055, "effective_prefill_tok_s": 885.4853310947822, "generation_ms": 1852.873400086537, "mean_inter_token_latency_ms": 3.6259753426351016, "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": 4123.1625999789685, "tpot_ms": 3.6259753426351016, "ttft_ms": 4123.16339998506, "visible_content_truncated": false, "visible_generation_ms": 1852.8726000804454 }, { "content_chunks": 45, "decode_tok_s": 218.16628377015783, "e2e_ms": 6162.258299998939, "effective_prefill_tok_s": 956.0188902475583, "generation_ms": 2342.2501001041383, "mean_inter_token_latency_ms": 4.583659687092247, "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": 3820.008199894801, "tpot_ms": 4.583659687092247, "ttft_ms": 3820.0087998993695, "visible_content_truncated": false, "visible_generation_ms": 2342.2495000995696 }, { "content_chunks": 45, "decode_tok_s": 98.81453306989384, "e2e_ms": 11971.510399947874, "effective_prefill_tok_s": 537.1895849698927, "generation_ms": 5171.304099960253, "mean_inter_token_latency_ms": 10.119968884462335, "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": 6800.206299987622, "tpot_ms": 10.119968884462335, "ttft_ms": 6800.206799991429, "visible_content_truncated": false, "visible_generation_ms": 5171.303599956445 }, { "content_chunks": 45, "decode_tok_s": 112.63117398830173, "e2e_ms": 6039.844100014307, "effective_prefill_tok_s": 2430.614861824142, "generation_ms": 4536.93220007699, "mean_inter_token_latency_ms": 8.878536595062602, "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": 1502.9118999373168, "tpot_ms": 8.878536595062602, "ttft_ms": 1502.9122999403626, "visible_content_truncated": false, "visible_generation_ms": 4536.931800073944 }, { "content_chunks": 45, "decode_tok_s": 120.75605674271378, "e2e_ms": 6038.9926000498235, "effective_prefill_tok_s": 2021.777207424257, "generation_ms": 4231.6718000220135, "mean_inter_token_latency_ms": 8.281158121373803, "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": 1807.32080002781, "tpot_ms": 8.281158121373803, "ttft_ms": 1807.3212000308558, "visible_content_truncated": false, "visible_generation_ms": 4231.671400018968 }, { "content_chunks": 45, "decode_tok_s": 218.39340917991672, "e2e_ms": 6038.265100098215, "effective_prefill_tok_s": 987.7108277869014, "generation_ms": 2339.81420006603, "mean_inter_token_latency_ms": 4.578892759424717, "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": 3698.450900032185, "tpot_ms": 4.578892759424717, "ttft_ms": 3698.4514000359923, "visible_content_truncated": false, "visible_generation_ms": 2339.813700062223 }, { "content_chunks": 45, "decode_tok_s": 181.39074378062028, "e2e_ms": 6037.853199988604, "effective_prefill_tok_s": 1134.525261701343, "generation_ms": 2817.1227999264374, "mean_inter_token_latency_ms": 5.512960469523361, "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": 3220.730400062166, "tpot_ms": 5.512960469523361, "ttft_ms": 3220.730800065212, "visible_content_truncated": false, "visible_generation_ms": 2817.1223999233916 }, { "content_chunks": 44, "decode_tok_s": 149.74379598161295, "e2e_ms": 5995.6129000056535, "effective_prefill_tok_s": 1414.9568722976994, "generation_ms": 3412.495300057344, "mean_inter_token_latency_ms": 6.678072994241378, "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": 2583.1175999483094, "tpot_ms": 6.678072994241378, "ttft_ms": 2583.1180999521166, "visible_content_truncated": false, "visible_generation_ms": 3412.494800053537 }, { "content_chunks": 45, "decode_tok_s": 268.0225355022769, "e2e_ms": 5851.803900091909, "effective_prefill_tok_s": 926.431034298317, "generation_ms": 1906.55610000249, "mean_inter_token_latency_ms": 3.7310295499070256, "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": 3945.247800089419, "tpot_ms": 3.7310295499070256, "ttft_ms": 3945.2484000939876, "visible_content_truncated": false, "visible_generation_ms": 1906.5554999979213 }, { "content_chunks": 44, "decode_tok_s": 99.29372046235044, "e2e_ms": 11816.92880007904, "effective_prefill_tok_s": 547.928267476314, "generation_ms": 5146.347600035369, "mean_inter_token_latency_ms": 10.071130332750233, "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": 6670.581200043671, "tpot_ms": 10.071130332750233, "ttft_ms": 6670.5818000482395, "visible_content_truncated": false, "visible_generation_ms": 5146.347000030801 }, { "content_chunks": 45, "decode_tok_s": 148.6487897698524, "e2e_ms": 6051.637300057337, "effective_prefill_tok_s": 1398.2379982536822, "generation_ms": 3437.6331000821665, "mean_inter_token_latency_ms": 6.727266340669602, "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": 2614.00419997517, "tpot_ms": 6.727266340669602, "ttft_ms": 2614.004599978216, "visible_content_truncated": false, "visible_generation_ms": 3437.6327000791207 }, { "content_chunks": 45, "decode_tok_s": 112.24901690958714, "e2e_ms": 6051.192100043409, "effective_prefill_tok_s": 2438.5952703025077, "generation_ms": 4552.378399996087, "mean_inter_token_latency_ms": 8.908763992164554, "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": 1498.8137000473216, "tpot_ms": 8.908763992164554, "ttft_ms": 1498.8142000511289, "visible_content_truncated": false, "visible_generation_ms": 4552.37789999228 }, { "content_chunks": 44, "decode_tok_s": 126.83358926611092, "e2e_ms": 6007.502100081183, "effective_prefill_tok_s": 1847.2649031593664, "generation_ms": 4028.9012000430375, "mean_inter_token_latency_ms": 7.884346771121404, "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": 1978.6009000381455, "tpot_ms": 7.884346771121404, "ttft_ms": 1978.6015000427142, "visible_content_truncated": false, "visible_generation_ms": 4028.900600038469 }, { "content_chunks": 45, "decode_tok_s": 164.11200056521636, "e2e_ms": 6050.268999999389, "effective_prefill_tok_s": 1245.0018589732802, "generation_ms": 3113.7271999614313, "mean_inter_token_latency_ms": 6.0933996085350906, "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": 2936.5418000379577, "tpot_ms": 6.0933996085350906, "ttft_ms": 2936.5424000425264, "visible_content_truncated": false, "visible_generation_ms": 3113.7265999568626 }, { "content_chunks": 45, "decode_tok_s": 206.17251459727538, "e2e_ms": 6197.488599922508, "effective_prefill_tok_s": 983.3336905051865, "generation_ms": 2478.5068999044597, "mean_inter_token_latency_ms": 4.850307044822817, "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": 3718.9817000180483, "tpot_ms": 4.850307044822817, "ttft_ms": 3718.98219990544, "visible_content_truncated": false, "visible_generation_ms": 2478.506400017068 }, { "content_chunks": 44, "decode_tok_s": 99.68356029282783, "e2e_ms": 11964.865899994038, "effective_prefill_tok_s": 534.901324377411, "generation_ms": 5126.2213999871165, "mean_inter_token_latency_ms": 10.031744422675375, "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": 6838.644500006922, "tpot_ms": 10.031744422675375, "ttft_ms": 6838.644800009206, "visible_content_truncated": false, "visible_generation_ms": 5126.221099984832 }, { "content_chunks": 45, "decode_tok_s": 250.58030671724805, "e2e_ms": 6196.178599959239, "effective_prefill_tok_s": 880.2206599577397, "generation_ms": 2039.266400039196, "mean_inter_token_latency_ms": 3.990736594988642, "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": 4156.912199920043, "tpot_ms": 3.990736594988642, "ttft_ms": 4156.912699923851, "visible_content_truncated": false, "visible_generation_ms": 2039.2659000353888 }, { "content_chunks": 45, "decode_tok_s": 112.80951614415031, "e2e_ms": 6045.075600035489, "effective_prefill_tok_s": 2415.3379501792374, "generation_ms": 4529.759699944407, "mean_inter_token_latency_ms": 8.86450039128064, "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": 1515.3159000910819, "tpot_ms": 8.86450039128064, "ttft_ms": 1515.3163000941277, "visible_content_truncated": false, "visible_generation_ms": 4529.759299941361 }, { "content_chunks": 45, "decode_tok_s": 121.1367022454455, "e2e_ms": 6045.111300074495, "effective_prefill_tok_s": 2004.1203531785845, "generation_ms": 4218.374700052664, "mean_inter_token_latency_ms": 8.255136399320282, "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": 1826.7366000218317, "tpot_ms": 8.255136399320282, "ttft_ms": 1826.7370000248775, "visible_content_truncated": false, "visible_generation_ms": 4218.374300049618 }, { "content_chunks": 44, "decode_tok_s": 184.5023246583656, "e2e_ms": 6002.1768000442535, "effective_prefill_tok_s": 1132.537515089569, "generation_ms": 2769.6127999806777, "mean_inter_token_latency_ms": 5.419985909942618, "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": 3232.564000063576, "tpot_ms": 5.419985909942618, "ttft_ms": 3232.564500067383, "visible_content_truncated": false, "visible_generation_ms": 2769.6122999768704 }, { "content_chunks": 44, "decode_tok_s": 223.28874520392975, "e2e_ms": 6000.549199990928, "effective_prefill_tok_s": 986.521508470235, "generation_ms": 2288.5166000341997, "mean_inter_token_latency_ms": 4.47850606660313, "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": 3712.0325999567285, "tpot_ms": 4.47850606660313, "ttft_ms": 3712.0329999597743, "visible_content_truncated": false, "visible_generation_ms": 2288.516200031154 }, { "content_chunks": 44, "decode_tok_s": 150.16954758710438, "e2e_ms": 6000.898600090295, "effective_prefill_tok_s": 1409.8882781833765, "generation_ms": 3402.820400078781, "mean_inter_token_latency_ms": 6.659139726181567, "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": 2598.078200011514, "tpot_ms": 6.659139726181567, "ttft_ms": 2598.07860001456, "visible_content_truncated": false, "visible_generation_ms": 3402.820000075735 }, { "content_chunks": 44, "decode_tok_s": 275.8869861050338, "e2e_ms": 5813.567399978638, "effective_prefill_tok_s": 925.1873794852406, "generation_ms": 1852.2077000234276, "mean_inter_token_latency_ms": 3.6246726027855725, "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": 3961.35969995521, "tpot_ms": 3.6246726027855725, "ttft_ms": 3961.360299959779, "visible_content_truncated": false, "visible_generation_ms": 1852.207100018859 }, { "content_chunks": 45, "decode_tok_s": 97.82246990566618, "e2e_ms": 11859.801799990237, "effective_prefill_tok_s": 552.2861171804814, "generation_ms": 5223.748699994758, "mean_inter_token_latency_ms": 10.222600195684457, "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": 6636.053099995479, "tpot_ms": 10.222600195684457, "ttft_ms": 6636.053599999286, "visible_content_truncated": false, "visible_generation_ms": 5223.7481999909505 }, { "content_chunks": 44, "decode_tok_s": 107.59928703082967, "e2e_ms": 6047.345399972983, "effective_prefill_tok_s": 2823.815844198928, "generation_ms": 4749.10210003145, "mean_inter_token_latency_ms": 9.293741878730822, "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": 1298.2432999415323, "tpot_ms": 9.293741878730822, "ttft_ms": 1298.2435999438167, "visible_content_truncated": false, "visible_generation_ms": 4749.101800029166 }, { "content_chunks": 44, "decode_tok_s": 128.41970232814293, "e2e_ms": 6048.40690002311, "effective_prefill_tok_s": 1772.12536203138, "generation_ms": 3979.14019995369, "mean_inter_token_latency_ms": 7.786967123197045, "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": 2069.26670006942, "tpot_ms": 7.786967123197045, "ttft_ms": 2069.267100072466, "visible_content_truncated": false, "visible_generation_ms": 3979.1397999506444 }, { "content_chunks": 44, "decode_tok_s": 153.01511798569496, "e2e_ms": 6047.553399926983, "effective_prefill_tok_s": 1353.7595686941536, "generation_ms": 3339.5392999518663, "mean_inter_token_latency_ms": 6.535301956852967, "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": 2708.0140999751166, "tpot_ms": 6.535301956852967, "ttft_ms": 2708.0144999781623, "visible_content_truncated": false, "visible_generation_ms": 3339.5388999488205 }, { "content_chunks": 44, "decode_tok_s": 178.79066206095675, "e2e_ms": 6046.511999913491, "effective_prefill_tok_s": 1150.0990976763853, "generation_ms": 2858.091100002639, "mean_inter_token_latency_ms": 5.593133268106926, "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": 3188.4208999108523, "tpot_ms": 5.593133268106926, "ttft_ms": 3188.4211999131367, "visible_content_truncated": false, "visible_generation_ms": 2858.0908000003546 }, { "content_chunks": 44, "decode_tok_s": 229.51794851945849, "e2e_ms": 6047.412700019777, "effective_prefill_tok_s": 959.9562921698605, "generation_ms": 2226.4054000843316, "mean_inter_token_latency_ms": 4.356957730106324, "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": 3821.007299935445, "tpot_ms": 4.356957730106324, "ttft_ms": 3821.007899940014, "visible_content_truncated": false, "visible_generation_ms": 2226.404800079763 }, { "content_chunks": 44, "decode_tok_s": 270.2779467463737, "e2e_ms": 6046.960200066678, "effective_prefill_tok_s": 882.7533454538999, "generation_ms": 1890.6463000457734, "mean_inter_token_latency_ms": 3.699894912026954, "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": 4156.313900020905, "tpot_ms": 3.699894912026954, "ttft_ms": 4156.314400024712, "visible_content_truncated": false, "visible_generation_ms": 1890.6458000419661 }, { "content_chunks": 45, "decode_tok_s": 98.30358976271981, "e2e_ms": 12069.40390006639, "effective_prefill_tok_s": 533.9662028659659, "generation_ms": 5198.182500083931, "mean_inter_token_latency_ms": 10.172568493314934, "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": 6871.22139998246, "tpot_ms": 10.172568493314934, "ttft_ms": 6871.221899986267, "visible_content_truncated": false, "visible_generation_ms": 5198.1820000801235 }, { "content_chunks": 45, "decode_tok_s": 107.31115295018708, "e2e_ms": 6069.034400046803, "effective_prefill_tok_s": 2807.568776999536, "generation_ms": 4761.853600037284, "mean_inter_token_latency_ms": 9.318695890483921, "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": 1307.1808000095189, "tpot_ms": 9.318695890483921, "ttft_ms": 1307.1814000140876, "visible_content_truncated": false, "visible_generation_ms": 4761.8530000327155 }, { "content_chunks": 45, "decode_tok_s": 152.45210863054814, "e2e_ms": 6070.977000053972, "effective_prefill_tok_s": 1350.0767366436944, "generation_ms": 3351.8723000306636, "mean_inter_token_latency_ms": 6.559436986361377, "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": 2719.1047000233084, "tpot_ms": 6.559436986361377, "ttft_ms": 2719.1052000271156, "visible_content_truncated": false, "visible_generation_ms": 3351.8718000268564 }, { "content_chunks": 45, "decode_tok_s": 268.5236240208317, "e2e_ms": 6070.680600008927, "effective_prefill_tok_s": 881.0652385671647, "generation_ms": 1902.9982999200001, "mean_inter_token_latency_ms": 3.7240671231311158, "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": 4167.682300088927, "tpot_ms": 3.7240671231311158, "ttft_ms": 4167.6828999770805, "visible_content_truncated": false, "visible_generation_ms": 1902.9977000318468 }, { "content_chunks": 44, "decode_tok_s": 129.55176711030649, "e2e_ms": 6024.008800042793, "effective_prefill_tok_s": 1765.6905552222784, "generation_ms": 3944.3692000349984, "mean_inter_token_latency_ms": 7.718922113571425, "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": 2079.639600007795, "tpot_ms": 7.718922113571425, "ttft_ms": 2079.6402000123635, "visible_content_truncated": false, "visible_generation_ms": 3944.3686000304297 }, { "content_chunks": 45, "decode_tok_s": 177.7469389369701, "e2e_ms": 6069.417299935594, "effective_prefill_tok_s": 1149.4599729485933, "generation_ms": 2874.873699969612, "mean_inter_token_latency_ms": 5.625975929490434, "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": 3194.543599965982, "tpot_ms": 5.625975929490434, "ttft_ms": 3194.543999969028, "visible_content_truncated": false, "visible_generation_ms": 2874.8732999665663 }, { "content_chunks": 45, "decode_tok_s": 98.44023850179312, "e2e_ms": 12081.020699930377, "effective_prefill_tok_s": 532.9421220822927, "generation_ms": 5190.966699970886, "mean_inter_token_latency_ms": 10.158447553759073, "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": 6890.053999959491, "tpot_ms": 10.158447553759073, "ttft_ms": 6890.054399962537, "visible_content_truncated": false, "visible_generation_ms": 5190.96629996784 }, { "content_chunks": 45, "decode_tok_s": 228.55898657072302, "e2e_ms": 6069.3294000811875, "effective_prefill_tok_s": 957.8507332064436, "generation_ms": 2235.746700083837, "mean_inter_token_latency_ms": 4.375238160633732, "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": 3833.5826999973506, "tpot_ms": 4.375238160633732, "ttft_ms": 3833.583200001158, "visible_content_truncated": false, "visible_generation_ms": 2235.7462000800297 }, { "content_chunks": 45, "decode_tok_s": 125.25419125052673, "e2e_ms": 6058.868899941444, "effective_prefill_tok_s": 1855.3277844026943, "generation_ms": 4079.7037999145687, "mean_inter_token_latency_ms": 7.983764774783892, "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": 1979.1651000268757, "tpot_ms": 7.983764774783892, "ttft_ms": 1979.165600030683, "visible_content_truncated": false, "visible_generation_ms": 4079.7032999107614 }, { "content_chunks": 45, "decode_tok_s": 140.4129487114166, "e2e_ms": 6057.793500018306, "effective_prefill_tok_s": 1519.1058362714653, "generation_ms": 3639.26550000906, "mean_inter_token_latency_ms": 7.121850293559804, "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": 2418.528000009246, "tpot_ms": 7.121850293559804, "ttft_ms": 2418.5285000130534, "visible_content_truncated": false, "visible_generation_ms": 3639.2650000052527 }, { "content_chunks": 44, "decode_tok_s": 113.38678814009911, "e2e_ms": 6011.087900027633, "effective_prefill_tok_s": 2442.185869329016, "generation_ms": 4506.697900011204, "mean_inter_token_latency_ms": 8.819369667340908, "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": 1504.3900000164285, "tpot_ms": 8.819369667340908, "ttft_ms": 1504.3904000194743, "visible_content_truncated": false, "visible_generation_ms": 4506.697500008158 }, { "content_chunks": 45, "decode_tok_s": 250.09279373473225, "e2e_ms": 6199.464499950409, "effective_prefill_tok_s": 884.2162916663078, "generation_ms": 2043.241599923931, "mean_inter_token_latency_ms": 3.9985158511231527, "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": 4156.222900026478, "tpot_ms": 3.9985158511231527, "ttft_ms": 4156.223900034092, "visible_content_truncated": false, "visible_generation_ms": 2043.2405999163166 }, { "content_chunks": 45, "decode_tok_s": 170.5852934968753, "e2e_ms": 6198.819200042635, "effective_prefill_tok_s": 1147.5843770228837, "generation_ms": 2995.5689000198618, "mean_inter_token_latency_ms": 5.862170058747283, "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": 3203.2503000227734, "tpot_ms": 5.862170058747283, "ttft_ms": 3203.250900027342, "visible_content_truncated": false, "visible_generation_ms": 2995.568300015293 }, { "content_chunks": 44, "decode_tok_s": 234.99001291623352, "e2e_ms": 6011.7964000673965, "effective_prefill_tok_s": 958.2418427827914, "generation_ms": 2174.5605000760406, "mean_inter_token_latency_ms": 4.255500000148808, "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": 3837.235899991356, "tpot_ms": 4.255500000148808, "ttft_ms": 3837.2364999959245, "visible_content_truncated": false, "visible_generation_ms": 2174.559900071472 }, { "content_chunks": 45, "decode_tok_s": 97.91205512528263, "e2e_ms": 12053.408599924296, "effective_prefill_tok_s": 538.0104767626195, "generation_ms": 5218.969199922867, "mean_inter_token_latency_ms": 10.213246966580954, "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": 6834.439400001429, "tpot_ms": 10.213246966580954, "ttft_ms": 6834.4397998880595, "visible_content_truncated": false, "visible_generation_ms": 5218.968800036237 }, { "content_chunks": 45, "decode_tok_s": 106.72935474107135, "e2e_ms": 6088.279499905184, "effective_prefill_tok_s": 2828.211960506214, "generation_ms": 4787.811200018041, "mean_inter_token_latency_ms": 9.36949354210967, "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": 1300.468299887143, "tpot_ms": 9.36949354210967, "ttft_ms": 1300.4688998917118, "visible_content_truncated": false, "visible_generation_ms": 4787.810600013472 }, { "content_chunks": 44, "decode_tok_s": 128.90787957879024, "e2e_ms": 6041.53240006417, "effective_prefill_tok_s": 1770.9114484484649, "generation_ms": 3964.0710999956354, "mean_inter_token_latency_ms": 7.757477690793807, "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": 2077.461300068535, "tpot_ms": 7.757477690793807, "ttft_ms": 2077.461800072342, "visible_content_truncated": false, "visible_generation_ms": 3964.070599991828 }, { "content_chunks": 44, "decode_tok_s": 179.4312891961777, "e2e_ms": 6040.785299963318, "effective_prefill_tok_s": 1152.5577778266047, "generation_ms": 2847.886799951084, "mean_inter_token_latency_ms": 5.573163992076485, "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": 3192.898500012234, "tpot_ms": 5.573163992076485, "ttft_ms": 3192.8989000152797, "visible_content_truncated": false, "visible_generation_ms": 2847.886399948038 }, { "content_chunks": 44, "decode_tok_s": 153.8051823468015, "e2e_ms": 6041.374100022949, "effective_prefill_tok_s": 1353.0763067198518, "generation_ms": 3322.384800063446, "mean_inter_token_latency_ms": 6.501731506973476, "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": 2718.989299959503, "tpot_ms": 6.501731506973476, "ttft_ms": 2718.990199966356, "visible_content_truncated": false, "visible_generation_ms": 3322.383900056593 }, { "content_chunks": 45, "decode_tok_s": 226.0783250312745, "e2e_ms": 6092.050399980508, "effective_prefill_tok_s": 960.3912216412679, "generation_ms": 2260.2786000352353, "mean_inter_token_latency_ms": 4.423245792632555, "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": 3831.7717999452725, "tpot_ms": 4.423245792632555, "ttft_ms": 3831.7722999490798, "visible_content_truncated": false, "visible_generation_ms": 2260.278100031428 }, { "content_chunks": 45, "decode_tok_s": 98.71405909984895, "e2e_ms": 11900.089200004004, "effective_prefill_tok_s": 547.4809510530278, "generation_ms": 5176.567599992268, "mean_inter_token_latency_ms": 10.130269275914419, "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": 6723.521600011736, "tpot_ms": 10.130269275914419, "ttft_ms": 6723.521999898367, "visible_content_truncated": false, "visible_generation_ms": 5176.567200105637 }, { "content_chunks": 45, "decode_tok_s": 265.8330779623592, "e2e_ms": 5904.718299978413, "effective_prefill_tok_s": 924.5543093105181, "generation_ms": 1922.2588998964056, "mean_inter_token_latency_ms": 3.7617590996015764, "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": 3982.459400082007, "tpot_ms": 3.7617590996015764, "ttft_ms": 3982.459800085053, "visible_content_truncated": false, "visible_generation_ms": 1922.2584998933598 }, { "content_chunks": 45, "decode_tok_s": 205.3865207560286, "e2e_ms": 6046.46009998396, "effective_prefill_tok_s": 1034.9959007545144, "generation_ms": 2487.99189995043, "mean_inter_token_latency_ms": 4.868868688748395, "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": 3558.4682000335306, "tpot_ms": 4.868868688748395, "ttft_ms": 3558.468700037338, "visible_content_truncated": false, "visible_generation_ms": 2487.9913999466226 }, { "content_chunks": 44, "decode_tok_s": 127.37635737595933, "e2e_ms": 6000.800400041044, "effective_prefill_tok_s": 1852.1247324683716, "generation_ms": 4011.733500054106, "mean_inter_token_latency_ms": 7.850750489342674, "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": 1989.0668999869376, "tpot_ms": 7.850750489342674, "ttft_ms": 1989.0672999899834, "visible_content_truncated": false, "visible_generation_ms": 4011.7331000510603 }, { "content_chunks": 45, "decode_tok_s": 174.0353613290382, "e2e_ms": 6045.472399913706, "effective_prefill_tok_s": 1185.1589793802536, "generation_ms": 2936.1848999978974, "mean_inter_token_latency_ms": 5.745958708410758, "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": 3109.2874999158084, "tpot_ms": 5.745958708410758, "ttft_ms": 3109.2881999211386, "visible_content_truncated": false, "visible_generation_ms": 2936.1841999925673 }, { "content_chunks": 45, "decode_tok_s": 149.6880878103215, "e2e_ms": 6045.576700009406, "effective_prefill_tok_s": 1400.5562860918733, "generation_ms": 3413.765300065279, "mean_inter_token_latency_ms": 6.680558317153188, "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": 2631.8113999441266, "tpot_ms": 6.680558317153188, "ttft_ms": 2631.8117999471724, "visible_content_truncated": false, "visible_generation_ms": 3413.764900062233 }, { "content_chunks": 44, "decode_tok_s": 113.7398393705312, "e2e_ms": 6000.119099975564, "effective_prefill_tok_s": 2445.253617506705, "generation_ms": 4492.709000012837, "mean_inter_token_latency_ms": 8.791994129183635, "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": 1507.4100999627262, "tpot_ms": 8.791994129183635, "ttft_ms": 1507.4105999665335, "visible_content_truncated": false, "visible_generation_ms": 4492.70850000903 }, { "content_chunks": 44, "decode_tok_s": 276.27865520180313, "e2e_ms": 5995.130599942058, "effective_prefill_tok_s": 889.3876943225606, "generation_ms": 1849.581899936311, "mean_inter_token_latency_ms": 3.6195340507559903, "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": 4145.548700005747, "tpot_ms": 3.6195340507559903, "ttft_ms": 4145.549100008793, "visible_content_truncated": false, "visible_generation_ms": 1849.5814999332651 }, { "content_chunks": 45, "decode_tok_s": 274.74039989645, "e2e_ms": 8674.964000005275, "effective_prefill_tok_s": 541.3038458639646, "generation_ms": 1859.937599976547, "mean_inter_token_latency_ms": 3.6397996085646716, "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": 6815.026400028728, "tpot_ms": 3.6397996085646716, "ttft_ms": 6815.027000033297, "visible_content_truncated": false, "visible_generation_ms": 1859.9369999719784 }, { "content_chunks": 45, "decode_tok_s": 356.2181041683724, "e2e_ms": 2725.225300062448, "effective_prefill_tok_s": 2858.114857525705, "generation_ms": 1434.5144000835717, "mean_inter_token_latency_ms": 2.807268884703663, "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": 1290.7108999788761, "tpot_ms": 2.807268884703663, "ttft_ms": 1290.7111999811605, "visible_content_truncated": false, "visible_generation_ms": 1434.5141000812873 } ], "requests": 100, "response_words": 256, "run_id": "benchmark-20260905T041029Z", "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": "2026-09-05T04:10:29Z", "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": 87084.14019993506 }