← AI Hashrate 中文

RTX 4090D 24GB (China) vs Titan X Pascal 12GB for Local LLMs

The RTX 4090D 24GB (China) has about 110% more memory bandwidth (1,008 vs 480 GB/s). Batch-1 decode tracks bandwidth closely, so a model that fits both cards should decode roughly 110% faster on the RTX 4090D 24GB (China). The RTX 4090D 24GB (China) fits more of the catalog at Q4/8K — 35 vs 18 of 69 models (24 vs 12 GB VRAM). Largest model that fits the RTX 4090D 24GB (China) but not the Titan X Pascal 12GB: Qwen3.6-35B-A3B (36.0B). MSRP is $1,599 for the RTX 4090D 24GB (China) vs $1,200 for the Titan X Pascal 12GB. MSRP is a launch list price, not live retail — Amazon prices move; click through for the current price. Not retail-shoppable: RTX 4090D 24GB (China) (datacenter-class; cloud rental is the realistic way to use it). In the Q4/8K head-to-head below (41 shared models), the RTX 4090D 24GB (China) posts the higher tok/s in 37 rows vs 4. Methodology.

Spec comparison

RTX 4090D 24GB (China)Titan X Pascal 12GB
VendorNVIDIANVIDIA
VRAM24 GB12 GB
Memory typeGDDR6XGDDR5X
Memory bandwidth1,008 GB/s480 GB/s
TDP425 W250 W
MSRP (list)$1,599$1,200
BuySearch Amazon (affiliate)

MSRP is the launch list price, not live retail. Buy links are Amazon affiliate links — the price after you click is the live price.

Q4 / 8K head-to-head — top 15 of 41 shared models

ModelParams (B)RTX 4090D 24GB (China) tok/sRTX 4090D 24GB (China) fitsTitan X Pascal 12GB tok/sTitan X Pascal 12GB fitsWinner
MiniCPM5-1B1.08593.9 est.Yes282.8 est.YesRTX 4090D 24GB (China)
Qwen3.5-2B2.0320.7 est.Yes152.7 est.YesRTX 4090D 24GB (China)
DeepSeek-R1-0528-Qwen3-8B-layer-mix-bpw-3.8-mlx2.3278.9 est.Yes132.8 est.YesRTX 4090D 24GB (China)
Qwen3.6-35B-A3B36.0213.8 est.Yes95.7 est.NoRTX 4090D 24GB (China)
Qwen3.5-35B-A3B34.7213.8 est.Yes102.6 est.NoRTX 4090D 24GB (China)
GLM-4.7-Flash31.2213.8 est.Yes118.5 est.NoRTX 4090D 24GB (China)
Qwen3-30B-A3B-Instruct-250730.5213.8 est.Yes123.5 est.NoRTX 4090D 24GB (China)
Nemotron-3-Nano-Omni-30B-A3B-Reasoning-BF1630.0213.8 est.Yes135.8 est.NoRTX 4090D 24GB (China)
Ornith-1.0-35B35.0213.8 est.Yes96.0 est.NoRTX 4090D 24GB (China)
qwen35b-a3b-fable-sft-abliterated34.7213.8 est.Yes97.5 est.NoRTX 4090D 24GB (China)
Llama-3.2-3B-Instruct3.2200.5 est.Yes95.5 est.YesRTX 4090D 24GB (China)
gpt-oss-20b21.5178.2 est.Yes189.3 est.NoTitan X Pascal 12GB
Qwen3-Coder-Next79.7174.1 est.No20.7 est.NoRTX 4090D 24GB (China)
gemma-4-26B-A4B-it25.2168.8 est.Yes135.0 est.NoRTX 4090D 24GB (China)
Qwen3-Next-80B-A3B-Instruct81.3167.5 est.No19.9 est.NoRTX 4090D 24GB (China)

est. = bandwidth-based estimate; measured = curated real benchmark that overrides the estimate. Fits = weights + KV(8K) + 1 GB overhead ≤ 95% of VRAM.

Fit coverage at Q4 / 8K

RTX 4090D 24GB (China)Titan X Pascal 12GB
Catalog models that fit35 of 6918 of 69
Largest exclusive fitQwen3.6-35B-A3B (36.0B)

RTX 4090D 24GB (China) details → · Titan X Pascal 12GB details →