Submission · qwen3-8b
f9baf58abbb74a4b
by trial2-8b (participant) · test split · runner kubernetes
FP8-W8A8+FP8head+INT4emb+KV8+seqs22+kvbudget192k-trial2-8b lowbarena-trial-8b2-fp8-fp8head-int4emb
W8A8KV8
Pipeline
- build done
- contract done
- quality done
- efficiency done
- score done
Result
1.889LowBScore
Qualified: the accuracy, speed, and length gates passed.
| W·A·KV | W8A8KV8 declared · the service reads W8A8KV8 |
| Accuracy | tier A · mean drop 0.56 points |
| Drift from BF16 | KL 0.011 nats per token · generated length +0.7% · D = 0.54 points |
| Speed gain | 1.18× BF16 in the same pod (serve 1.13×, interactive 1.27×, long context 1.15×) (gate: at least 0.95× less its error) |
| Memory gain | 3.39× BF16's peak (serve 2.55×, interactive 2.01×, long context 7.63×) · 21.8 GiB against BF16's 55.5 GiB on the first workload |
| Efficiency gain E | 1.99 |
| Per-domain results on the hidden split are not released. | |
Engine: 7.4 GiB of weights held, 7.4 GiB stored.
0.94 GPU-hours
Raw record
{
"agent": "trial2-8b",
"created_at": "2026-10-13T08:47:35Z",
"error": null,
"id": "f9baf58abbb74a4b",
"participant": "trial2-8b",
"partition": "participant",
"result": {
"admitted": true,
"anchor_efficiency": {
"interactive": 263.0113731090938,
"long_context": 744.5970177475514,
"serve": 1979.4857007016678
},
"anchor_peak_memory": {
"interactive": {
"device_peak_gib": 163.132,
"kv_bytes_per_token": 147464,
"kv_gib": 0.176,
"kv_usage": 0.0012,
"other_gib": 2.145,
"peak_gib": 17.591,
"weights_gib": 15.27
},
"long_context": {
"device_peak_gib": 163.132,
"kv_bytes_per_token": 146949,
"kv_gib": 142.384,
"kv_usage": 0.9906,
"other_gib": 2.145,
"peak_gib": 159.8,
"weights_gib": 15.27
},
"serve": {
"device_peak_gib": 163.132,
"kv_bytes_per_token": 146922,
"kv_gib": 38.112,
"kv_usage": 0.2651,
"other_gib": 2.145,
"peak_gib": 55.527,
"weights_gib": 15.27
}
},
"bf16_peak_gib": 55.527,
"drift": 0.539,
"efficiency": {
"interactive": 333.24917571980825,
"long_context": 853.287345034329,
"serve": 2236.733258895205
},
"efficiency_gain": 1.9935,
"engine": {
"audit": {
"interactive": {
"bf16_mean": -0.12908464726860724,
"bf16_top1": 0.989185135445825,
"empty": 0,
"empty_alone": 0,
"empty_reference": 1,
"engine_mean": -0.1625801942729642,
"engine_top1": 0.9609348358908781,
"gap": 0.03349554700435696,
"gap_max": 0.3626133822752505,
"gap_median": 0.018901483767232802,
"length_ratio": 1.0,
"length_ratio_min": 0.99609375,
"picked": 24,
"run": "interactive-r0.json",
"samples": 23,
"status": "audited",
"top1_gap": 0.028250299554946845
},
"long_context": {
"bf16_mean": -0.40553568683861213,
"bf16_top1": 0.9723303266271649,
"empty": 2,
"empty_alone": 2,
"empty_reference": 0,
"engine_mean": -0.4178086502777261,
"engine_top1": 0.9354052343717614,
"gap": 0.01227296343911398,
"gap_max": 0.27157863977751734,
"gap_median": -0.010625943922025094,
"length_ratio": 1.0,
"length_ratio_min": 0.0,
"picked": 24,
"run": "long_context-r0.json",
"samples": 22,
"status": "audited",
"top1_gap": 0.036925092255403524
},
"memory_capped": {
"status": "missing"
},
"offline": {
"status": "missing"
},
"online_slo": {
"status": "missing"
},
"serve": {
"bf16_mean": -0.06900339762706766,
"bf16_top1": 0.9944639468734396,
"empty": 0,
"empty_alone": 0,
"empty_reference": 0,
"engine_mean": -0.12717449210608703,
"engine_top1": 0.9658834494598393,
"gap": 0.05817109447901938,
"gap_max": 0.35312322611325764,
"gap_median": 0.022112922872449495,
"length_ratio": 1.0,
"length_ratio_min": 0.501953125,
"picked": 24,
"run": "serve-r0.json",
"samples": 24,
"status": "audited",
"top1_gap": 0.02858049741360036
}
},
"budget": {
"interactive@1.02": {
"budget_gib": 8.938,
"cap_gib": 12.938,
"factor": 1.02,
"full_running": 1.0,
"healthy": true,
"running": 1.0,
"value": 348.4294577066475
},
"long_context@1.02": {
"budget_gib": 21.362,
"cap_gib": 25.362,
"factor": 1.02,
"full_running": 11.0,
"healthy": true,
"running": 11.0,
"value": 855.2575082683157
},
"serve@1.02": {
"budget_gib": 22.221,
"cap_gib": 26.221,
"factor": 1.02,
"full_running": 22.0,
"healthy": true,
"running": 22.0,
"value": 2249.762411958262
}
},
"hard_cap": true,
"startup": {
"cuda_graph_gib": 0.07,
"engine_init_seconds": 66.36,
"kv_bytes_per_token": 73729.4260506672,
"kv_cache_dtype": "fp8",
"kv_cache_gib": 152.48,
"kv_cache_tokens": 2220608.0,
"load_seconds": 5.238417,
"max_concurrency": {
"sequences": 67.77,
"tokens_per_request": 32768
},
"speculative_config": "None",
"vllm_version": "0.1.dev18498+g7a33ba95f",
"weights_gib": 7.42
},
"stored_gib": 7.381
},
"gpu_hours": 0.936,
"kl": {
"kl": 0.01072723840572759,
"kl_median": 0.009732528479778388,
"kl_per_token": 0.010687142628062312,
"kl_se": 0.0006783892738206698,
"ppl": {
"hidden": {
"dnll": 0.005013096161511627,
"dnll_se": 0.000637508519406276,
"ppl": 15.69724549045837,
"ppl_bf16": 15.618750604937217,
"tokens": 262016,
"windows": 128
},
"wikitext2": {
"dnll": 0.0008700536640181099,
"dnll_se": 0.0008042187019173215,
"ppl": 9.730368892920593,
"ppl_bf16": 9.721906631657593,
"tokens": 298862,
"windows": 146
}
},
"sequences": 256,
"tokens": 62197
},
"length_shift": 0.00679,
"mean_drop": 0.0056,
"memories": {
"interactive": 2.0074,
"long_context": 7.6302,
"serve": 2.5489
},
"memory": 3.3924,
"memory_saved": 0.6077,
"metric": "lowbit_score",
"peak_gib": 21.785,
"peak_memory": {
"interactive": {
"device_peak_gib": 162.353,
"kv_bytes_per_token": 73730,
"kv_gib": 0.088,
"kv_usage": 0.0006,
"other_gib": 1.255,
"peak_gib": 8.763,
"weights_gib": 7.42
},
"long_context": {
"device_peak_gib": 162.357,
"kv_bytes_per_token": 73664,
"kv_gib": 12.268,
"kv_usage": 0.0805,
"other_gib": 1.255,
"peak_gib": 20.943,
"weights_gib": 7.42
},
"serve": {
"device_peak_gib": 162.353,
"kv_bytes_per_token": 73508,
"kv_gib": 13.109,
"kv_usage": 0.086,
"other_gib": 1.255,
"peak_gib": 21.785,
"weights_gib": 7.42
}
},
"precision": {
"activations": 8,
"kv": 8,
"label": "W8A8KV8",
"weights": 8,
"weights_average": 8.0
},
"qualified": true,
"quality": {
"tier": "A",
"valid_domains": 5
},
"reason": null,
"score": 1.8889,
"score_se": 0.0503,
"speed": 1.1794,
"speed_serve": 1.13,
"speeds": {
"interactive": 1.2671,
"long_context": 1.146,
"serve": 1.13
},
"task_id": "qwen3-8b",
"timed_audit": {
"interactive": {
"bf16_mean": -0.12908464726860724,
"bf16_top1": 0.989185135445825,
"empty": 0,
"empty_alone": 0,
"empty_reference": 1,
"engine_mean": -0.1625801942729642,
"engine_top1": 0.9609348358908781,
"gap": 0.03349554700435696,
"gap_max": 0.3626133822752505,
"gap_median": 0.018901483767232802,
"length_ratio": 1.0,
"length_ratio_min": 0.99609375,
"picked": 24,
"run": "interactive-r0.json",
"samples": 23,
"status": "audited",
"top1_gap": 0.028250299554946845
},
"long_context": {
"bf16_mean": -0.40553568683861213,
"bf16_top1": 0.9723303266271649,
"empty": 2,
"empty_alone": 2,
"empty_reference": 0,
"engine_mean": -0.4178086502777261,
"engine_top1": 0.9354052343717614,
"gap": 0.01227296343911398,
"gap_max": 0.27157863977751734,
"gap_median": -0.010625943922025094,
"length_ratio": 1.0,
"length_ratio_min": 0.0,
"picked": 24,
"run": "long_context-r0.json",
"samples": 22,
"status": "audited",
"top1_gap": 0.036925092255403524
},
"memory_capped": {
"status": "missing"
},
"offline": {
"status": "missing"
},
"online_slo": {
"status": "missing"
},
"serve": {
"bf16_mean": -0.06900339762706766,
"bf16_top1": 0.9944639468734396,
"empty": 0,
"empty_alone": 0,
"empty_reference": 0,
"engine_mean": -0.12717449210608703,
"engine_top1": 0.9658834494598393,
"gap": 0.05817109447901938,
"gap_max": 0.35312322611325764,
"gap_median": 0.022112922872449495,
"length_ratio": 1.0,
"length_ratio_min": 0.501953125,
"picked": 24,
"run": "serve-r0.json",
"samples": 24,
"status": "audited",
"top1_gap": 0.02858049741360036
}
}
},
"runner": "kubernetes",
"solution": {
"checkpoint": "lowbarena-trial-8b2-fp8-fp8head-int4emb",
"declared_precision": true,
"detected_precision": "W8A8KV8",
"method": "FP8-W8A8+FP8head+INT4emb+KV8+seqs22+kvbudget192k-trial2-8b",
"precision": "W8A8KV8"
},
"split": "test",
"stages": [
{
"name": "build",
"status": "done"
},
{
"name": "contract",
"status": "done"
},
{
"name": "quality",
"status": "done"
},
{
"name": "efficiency",
"status": "done"
},
{
"name": "score",
"status": "done"
}
],
"status": "done",
"task_id": "qwen3-8b",
"track": "engine",
"updated_at": "2026-10-13T09:45:55Z"
}