Tencent Hunyuan Hy3 1-bit quantization: fits on one GPU, but is it usable?

Tencent released IQ1_M quants for the 295B Hy3 MoE model, fitting into a single 96GB GPU at ~85.5 GiB. Actual BPW is ~2.4, not 1-bit. With 64K q8 KV cache, only ~2 GiB headroom remains—official config is marked tight. Agent and coding abilities dip slightly vs BF16; multi-step reliability is untested. Community GGUF hits ~23–25 tok/s on M3 Max, but 8K prompt prefill takes ~200s. Worth experimenting if you already own the hardware, not a reason to buy it.

Why it matters: A hands-on breakdown of Tencent Hunyuan Hy3's quantized release, answering with real numbers how much VRAM remains, what speed looks like, and where precision drops after 'it fits on one GPU.' Hits all three HKR axes, but as a single technical review rather than an official la...