Half of 36 inference configurations sit on the Pareto frontier, and AWQ 4-bit misses the quality floor by 5.9%.

arXiv 2609.17863 measured 54 configurations of Qwen2.5-7B-Instruct on vLLM 0.12 across L4, A100 and H100, then calibrated a simulator reproducing them with cross-campaign drift under 1.5%. On the calibrated grid, 18 of 36 configurations reach the cost/quality/latency frontier, and combined optimizations reach it more often than single ones. Quality testing on 200 GSM8K questions reorders the winners: AWQ 4-bit cuts per-token latency to 0.34x baseline on L4 but loses 5.9% strict accuracy, narrowly missing a 95% quality floor, while flexible answer extraction recovers FP16 parity. The loss is formatting, not arithmetic, which is a different fix entirely.