Unverified70% confidenceBenchmarkExact time
On a single RTX PRO 6000, peak VRAM for Qwen3.8-27B-NVFP4 dropped from 72.9 GB to 40.2 GB, and per-step time was reduced by approximately 11%
2
Sources
70%
Confidence
Medium-term (~90 days)
Relevance
10/1/2026
First Seen
Valid until: 12/30/2026
Sources
Related Entities
Related Claims
UnverifiedA ~25 GB Q3 model running entirely within the RTX 4090's VRAM (bandwidth ~1 TB/s) theoretically achieves ~40 tokens per second, whereas running it in DDR5 memory (~80 GB/s) drops throughput to 1–3 tokens per second.78% similarUnverifiedThe FP16 version of Gemma 3 12B requires approximately 24 GB of VRAM, while the Q4_K_M version reduces this to approximately 7–8 GB, with inference speed improving by approximately 2–3x and structured output accuracy losing only approximately 2–5%.75% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/964982API
curl https://kongchang.com/api/v1/knowledge/claims/964982MCP
get_claim(id=964982)