Upload README.md with huggingface_hub

README.md

@@ -47,6 +47,7 @@ Results are **split by reasoning mode**: comparing a thinking-on (reasoning) mod
 | Qwable-5-27B-Coder | 26.90B | Q6_K | 87.9 | 97.1 | 95.5 | 97.0 | 90.9 | **93.7** |
 | Qwen3.8-27B | 27.32B | Q6_K¹⁰ | 85.3 | 96.7 | 94.3 | 97.5 | 94.5 | **93.7** |
 | Qwen3.8-27B | 27.32B | Q8_0¹⁰ | 85.2 | 96.8 | 94.4 | 97.4 | 94.5 | **93.7** |
+| Qwen3.8-27B | 27.32B | BF16¹² | 85.3 | 96.8 | 94.3 | 97.4 | 93.9 | **93.5** |
 | Qwen3.8-27B | 27.32B | UD-Q4_K_XL¹⁰ | 85.1 | 96.6 | 94.4 | 97.3 | 93.9 | **93.5** |
 | Qwen3.6-35B-A3B | 34.66B | UD-Q4_K_M | 85.0 | 95.7 | 93.3 | 96.7 | 95.7 | **93.3** |
 | Qwen3.8-27B | 27.32B | Q4_K_M¹⁰ | 85.0 | 96.8 | 94.3 | 97.1 | 92.7 | **93.2** |
@@ -94,6 +95,8 @@ Results are **split by reasoning mode**: comparing a thinking-on (reasoning) mod
 
 ¹¹ NVFP4 checkpoint (unsloth day-0 cut, 22.6GB) served by vLLM 0.25.1 on sm_120, measured 2026-08-17 — the only non-llama.cpp Qwen3.8 row; quality same-harness over HTTP, cross-stack for speed. Q6_K-sized but lands below Q4_K_M, HumanEval pays nearly the whole tax (89.6 vs 94.5). The lane's draw is the shipped MTP speculative head, which llama.cpp cannot run yet: 1.72–1.81x decode at every prompt depth tested (69 → 114–125 tok/s effective, held at 32k). [NVFP4 report](reports/qwen3-8-27b-nvfp4.md).
 
+¹² Full-precision BF16 reference (split GGUF, 54.7GB — 1.7x the card), quality-only via llama.cpp partial offload (`-ngl 40`), banked 2026-08-17→20 across three overnight windows. Lands mid-ladder at 93.5 (ties UD-Q4_K_XL, 0.1 under the Q6_K/Q8_0 pair) with GPQA-diamond 48.5 inside the 4-bit band's noise band — the ladder's 4-bit-and-up rungs were already at the full-precision ceiling, and the IQ2 floor's loss is now measured against a true reference. No speed rows by design (partial-offload decode is not comparable to resident rungs). [BF16 addendum](reports/qwen3-8-27b-quant-ladder.md).
+
 ### Thinking ON (reasoning · extended chain-of-thought)
 
 | Model | Params | Quant | MMLU | ARC-C | HellaSwag | GSM8K | HumanEval | q_avg |