Running Qwen3.8-Flash on RTX 3090: Quantization & Performance
crusaderky · reddit · 2026-08-28
A detailed report on running Qwen3.8-Flash-next on an RTX 3090 with 12GB VRAM. The author achieved 160 tok/s prefill and 16 tok/s decode using IQ4XS weights and kvarn5 KV cache. The post includes variant setups for 16GB/12GB cards and links to specific llama.cpp builds and deployment recipes.
More from Infra
- SK Hynix ties new US plant to looser chip equipment exports to China, sources say — pstAsiatech · 2026-10-15
- Perplexity's new embedding models: index with a 9B model, search with 0.6B — antoine_chaffin · 2026-10-15
- Unsloth ships GGUF quants for Nemotron 3.5: Bolt-74B runs in 24GB RAM — danielhanchen · 2026-10-15
- vLLM-Omni ships Day-0 serving for FLUX 3 Image Klein with BF16, FP8, NVFP4 support — vllm_project · 2026-10-15
- NVIDIA's Nemotron 3.5 Super VL lands training and inference support on Applied Compute — ypatil125 · 2026-10-15
- CoreWeave bond yield tops 16% as loans outlast customer contracts by ~2 years — tengyanAI · 2026-10-15