AnalysisAI ModelsJune 23, 2026

KV cache quantization mapping for Qwen3.6 and Gemma4

Reddit user crusaderky maps KLD of KV cache quantization for Qwen3.6-35B-A3B and Gemma4-E2B QAT. Findings: q8/q8 nearly free on both, q4/q4 usable on Qwen but catastrophic on Gemma.

1 source