REAP vs parent matched comparison

reports/qwen3-6-35b-a3b-ud-q6-k.md

@@ -0,0 +1,94 @@
+# Benchmark Report: Qwen3.6-35B-A3B (UD-Q6_K)
+
+**Date:** 2026-06-03  
+**Author:** WITCHEER  
+**Platform:** NVIDIA GeForce RTX 5090 Benchmark Rig (capsule)  
+
+---
+
+## Model
+
+| Field | Value |
+|-------|-------|
+| Model | Qwen3.6-35B-A3B |
+| Parameters | 34.66 B (moe (3b active)) |
+| Quantization | UD-Q6_K |
+| File size | 27.3 GiB |
+| Engine | llama.cpp (CUDA 12.8 (patched)) |
+
+## Hardware
+
+| Component | Spec |
+|-----------|------|
+| GPU | NVIDIA GeForce RTX 5090 |
+| CPU | AMD Ryzen 5 9600 |
+| RAM | 64GB DDR5-5600 |
+| OS | Ubuntu 26.04 LTS |
+| CUDA | 12.8 (patched) |
+
+---
+
+## Quality Benchmarks
+
+All benchmarks use generative evaluation via llama-server chat completions. Multiple-choice tasks (MMLU, ARC, HellaSwag) use letter extraction instead of loglikelihood scoring -- results are internally consistent for model comparison but absolute scores may differ from logprob-based evaluations by 5-15%.
+
+### Summary
+
+| Benchmark | Score | Metric |
+|-----------|------:|--------|
+| **MMLU** | **94.74%** | accuracy |
+| **ARC-Challenge** | **97.00%** | accuracy |
+| **HellaSwag** | **87.00%** | accuracy |
+| **HumanEval** | **98.00%** | pass@1 |
+| **GSM8K** | **92.00%** | exact_match |
+
+### MMLU Breakdown by Category
+
+| Category | Score | Correct / Total |
+|----------|------:|----------------:|
+| Stem | 100.00% | 36 / 36 |
+| Humanities | 87.50% | 21 / 24 |
+| Social Sciences | 96.15% | 25 / 26 |
+| Other | 92.86% | 26 / 28 |
+
+---
+
+## Speed Benchmarks
+
+Measured with `llama-bench`. All layers GPU-offloaded (`-ngl 99`).
+
+### Prompt Processing (tokens/s)
+
+| Context Length | Speed | +/-sigma |
+|---------------:|------:|---------:|
+| 128 | 3,048 | 35.2 |
+| 512 | 8,321 | 54.8 |
+| 2048 | 8,125 | 51.1 |
+| 4096 | 7,901 | 51.1 |
+| 8192 | 7,599 | 26.1 |
+| 16384 | 6,998 | 20.6 |
+
+### Generation (tokens/s)
+
+| Metric | Speed | +/-sigma |
+|--------|------:|---------:|
+| tg128 | 259.6 | 1.5 |
+
+---
+
+## Methodology
+
+### Evaluation Framework
+
+Custom generative evaluators built for this rig. All benchmarks run through llama-server's `/v1/chat/completions` endpoint.
+
+- **Scoring:** Generative evaluation (not loglikelihood)
+- **Thinking:** enabled
+- **MCQ scoring:** First valid letter extracted from response (A/B/C/D)
+- **Temperature:** 0 (deterministic)
+- **Max tokens:** 2,048
+- **GPU offload:** All layers (`-ngl 99`)
+
+---
+
+*Benchmarked by WITCHEER on the RTX 5090 Benchmark Rig. Source: [github.com/notwitcheer/llm-bench-rig/blob/main/reports/qwen3-6-35b-a3b-ud-q6-k.md](https://github.com/notwitcheer/llm-bench-rig/blob/main/reports/qwen3-6-35b-a3b-ud-q6-k.md). Dataset: [huggingface.co/datasets/witcheer/rtx-5090-benchmarks/blob/main/reports/qwen3-6-35b-a3b-ud-q6-k.md](https://huggingface.co/datasets/witcheer/rtx-5090-benchmarks/blob/main/reports/qwen3-6-35b-a3b-ud-q6-k.md).*