#3657 [Benchmark] RTX 5080 / Qwen3.8-Flash-Next / Q2_0 / strata-qwen38-thinking-v1-draft.1

Beschreibung

Standard benchmark report for RTX 5080 / Qwen3.8-Flash-Next / Q2_0, protocol `strata-qwen38-thinking-v1-draft.1` (draft).

**Run:** `20261021T213810Z-cf02f5de` — folder `bench/results/2026-10-21-standard-rtx-5080-cf02f5de/` (`README.md`, `results.json`, `outputs.jsonl`, `config.json`, `engine-log-excerpt.txt`, `provenance/`).

**Status:** complete. 21 of 21 required requests finished, 0 failed, 0 skipped; sanity 12/12 (instruction 4/4, retrieval 4/4, tool calls 4/4). Optional concurrency stage not run. Time after model load: 1 min 14 s.

**Setup:** Strata v0.2.2 (commit `e6fff4ad12908de74952abe53b2bade45a68aa0f`), engine 0.1.47 release build, Q2_0 pack (GSQ-RCO shards), context 131072, KV q4_0, conversation cache 13312 MiB, VRAM reserve 1024 MiB. Run against a live server via `strata bench run --suite standard-v1 --url http://127.0.0.1:8080`.

**Hardware:** NVIDIA GeForce RTX 5080 (15.9 GiB), AMD Ryzen 7 7800X3D, 94.9 GiB RAM, NVMe (INTEL SSDPED1D480GA), Windows 10.0.26300, gpu-driver 617.42, PCIe Gen4 x16.

**Known nulls:** model/source/tokenizer revision, runtime and background work are `null` with reason in `results.json` (live server does not expose them; revisions not pinned). Engine config options come from `config.json`, no separate engine startup log was captured — a sanitized log excerpt covering the load and expert-cache behavior is included instead.