Qwen3.8-27B on 12GB: How Far Can an RTX 5070 Really Push It?
A complete Qwen3.8-27B campaign on WumboJetsII covering quant selection, context limits, templates, reasoning, speculative decoding, coding, reliability, and final readiness.
/records
Technical Reports
Long-form technical reports, benchmarks, experiments, fit tests, baselines, and findings. Historical reports retain their original evidence context.
Technical Reports preserve the longer technical context behind WumboLabs work. For current per-model evidence, use Evaluations. For WELP, review rules, and publication boundaries, use Methodology.
Reports
12B Gemma-Based Practical Use Test
Gemmable 4 12B, Gemma 4 12B QAT Q4, and Gemma 4 12B UD-Q5 tested through LLMGauge on WumboJetsII.
Mellum2 Agent Backend Test
JetBrains Mellum2 Instruct and Thinking Q4_K_M tested through LLMGauge on WumboJetsII.
Local LLM Baseline
Historical baseline notes for local LLM testing on WumboJetsII.
Building WumboLabs
Notes on building the WumboLabs website as a public technical report.