{
  "id": 20738048,
  "title": "Three 'e-waste' Nvidia V100s hit 100 tokens per second on Qwen 3.6 35B in a $1,042 homelab rig",
  "url": "https://urgent.news/2026/11/19/three-e-waste-nvidia-v100s-hit-100-tokens-per-second-on-qwen-3-6-35b",
  "topic": "tech",
  "section": "Tech",
  "published": "2026-11-19T11:00:00.000Z",
  "source": {
    "name": "Tom's Hardware",
    "slug": "tom-s-hardware",
    "url": "https://www.tomshardware.com/tech-industry/artificial-intelligence/three-e-waste-nvidia-v100s-hit-100-tokens-per-second-on-qwen-3-6-35b-in-a-usd1-042-homelab-rig-decommissioned-gpus-outpace-a-usd5-399-m5-max-mac-studio-at-long-contexts"
  },
  "original_language": "en",
  "account": "Three decommissioned Nvidia V100 GPUs powered a $1,042 home lab rig to achieve nearly 100 tokens per second when running Qwen 3.6 35B, according to a benchmark conducted by a homelab builder. The smaller Qwen 3.6 27B model managed about 24 tokens per second on the same platform, while the V100s consumed about half the energy of the same number of older P100 GPUs for a given job. The builder noted that scaling decommissioned GPUs will only get you so far, as the performance of the V100s is limited by the context size. For context lengths of 4K, the V100 setup achieved 100.8 tokens per second, falling to 72.7 tokens per second at 128K. In comparison, a modern Mac Studio with 40 GPU cores and 128GB of memory managed 43.9 tokens per second at 128K. The Nvidia DGX Spark, priced at $4,700, and the M5 Max Mac Studio, priced at $5,399, also demonstrated weaker performance with the same model. The benchmark showed the V100 setup using less energy compared to older P100 GPUs and significantly less energy than newer H100 and B200 GPUs, making it a more energy-efficient option.",
  "summary": "A homelab builder runs Qwen 3.6 35B on three used Nvidia Tesla V100 GPUs at about 100 tokens per second, in a $1,042 build.",
  "key_points": [],
  "editors_take": null,
  "illustration": null,
  "coverage": {
    "outlets": 1,
    "also_reported_by": []
  },
  "ai_generated": true,
  "disclaimer": "Summaries, key points and the editor’s take are written by software from other outlets’ reporting and may contain errors — always check the linked original."
}