LordNeel releases GGUF quants for Tencent Hy3 and NVIDIA Nemotron-Labs-Audex

LordNeel has published GGUF quantization sets for Tencent's 295B MoE Hy3 model and NVIDIA's audio-capable 30B hybrid MoE Nemotron-Labs-Audex-30B-A3B, providing full benchmark data and reproducible metrics.

• Hy3 quants include Q6_K (226 GiB), Q4_K_M (167 GiB), Q3_K_L (143 GiB), and IQ2_M (90 GiB), with quality measured against BF16 reference logits on WikiText-2.
• Nemotron-Labs-Audex quants offer text-only GGUFs and full-vocab audiogen files, including MXFP4_MOE for smaller file sizes and higher throughput.
• Both repos contain charts, checksums, quant plans, and per-quant logs to ensure reproducibility without subjective quality claims.

The release provides users with transparent, data-driven options for running these large models locally, including specific hardware recommendations and known limitations like tensor fallbacks.