AINewsnow

Homelab builder runs Qwen 3.6 35B on three used Nvidia V100s at 100 tokens per second

A $1,042 homelab rig using three decommissioned Nvidia Tesla V100 GPUs serves Qwen 3.6 35B at about 100 tokens per second, outpacing a $5,399 M5 Max Mac Studio at long contexts.

Read the full story at DEV Community — Machine Learning ↗

Timeline · 3 reports

  1. 2026-11-19 11:00 · Tom's Hardware
    Three 'e-waste' Nvidia V100s hit 100 tokens per second on Qwen 3.6 35B in a $1,042 homelab rig — decommissioned GPUs outpace a $5,399 M5 Max Mac Studio at long contexts
  2. 2026-11-19 11:00 · Tom's Hardware
    Three 'e-waste' Nvidia V100s hit 100 tokens per second on Qwen 3.6 35B in a $1,042 homelab rig
  3. 2026-11-18 21:35 · DEV Community — Machine Learning
    How many tokens per second should an LLM serve on one GPU?

More stories

  1. How to Speed Up Qwen3.8-27B in llama.cpp (NVIDIA's Draft Model May Lose to MTP) — DEV Community — Machine Learning
  2. Microsoft CEO warns AI-driven bug discovery is raising the pace of tech outages — OpenAI News
  3. Nvidia Fallout Adds New Tension to US-China AI Race — Bloomberg AI
  4. AWS Adds Managed NVIDIA Dynamo Support to SageMaker HyperPod Inference — AWS Machine Learning Blog
  5. Chinese tech firms fear AI hits from US curbs after Nvidia fallout — Washington Post AI
  6. Anthropic Appears Among Market-Moving Names Across Recent Trading Sessions — Barron's
  7. US proposal to curb China's remote access to AI chips rattles Southeast Asia's data center boom — Ars Technica AI
  8. National security and AI leaders urge DOJ to investigate Nvidia over chip smuggling — NBC News

Read the Daily Brief: the day's most important AI stories →