DeepSeek V4 Flash 284B (13B active)

on NVIDIA RTX 5060 Ti 16GB · llama.cpp · 128,000 ctx

Sep 7, 2026

Throughput

11.0 t/s gen · 200.0 tokens/s prompt processing (prefill)Prompt processing measures input prefill speed; it does not tell you the time to first token.

VRAM reported

16 GB

Summary

User reports llama.cpp running on 4x RTX 5060 Ti 16GB with DDR4 3200 RAM in 4-channel. Setup uses -ub/-b at 4096.

View source ↗