How to Make Local AI Stupid Fast with DeepSeek V4 + MTP 🤯

Published May 29, 2026

Views 7.65K | Likes 152 | Comments 49

Views#1,692,859

Views#14,866

Channel

xCreate

26.1K subs•263 videos•Feb 2012

Views

7,652

Likes

152

Comments

49

Performance vs. channel

Outlier score

Engagement rate

Revenue estimate

Total (all-time): $5 – $15

Based on estimated $0.7–$2 CPM.

More features coming

We're working on new analytics and tools. Stay tuned.

About

We saw the speed up MTP gave Gemma and Qwen, but now it's time for the big dog. How much faster will it make it go woof? TEST SYSTEM Inferencer App v1.11.7: https://inferencer.com M3 Ultra 512GB DeepSeek V4 Flash: https://huggingface.co/models?search=inferencerlabs/deepseek+v4+flash BUY NOW Mac Studio: https://vtudio…

Published: May 29, 2026
Made for kids: No

More features coming

We're working on new analytics and tools. Stay tuned.