Benchmark results for AEON-7/Qwen3.6-35B-A3B-heretic-NVFP4 + DFlash (NVIDIA DGX Spark, GB10)
The previous post rejected MTP-1/2 and projected ~2× from NVFP4. We measured +152% aggregate at c=10 and discovered that published benchmark numbers are only honest when paired with the workload that produced them.