Yhn HealthFitness first seen 17 h ago, last 1 h ago, peak #9
Qwen3.8 hits 255 tokens per second on one RTX 5090
Original: Running Qwen3.8 at 255 tok/s on a single 5090
A developer reports running the Qwen3.8 model at 255 tokens per second on a single Nvidia RTX 5090 GPU. The result has drawn attention on Hacker News, where readers are discussing the performance figure and what it suggests about inference efficiency on consumer hardware.
Why now: Achieving that inference speed for a modern language model on one consumer GPU is a striking benchmark result that invites discussion of optimization techniques.
Qwen3.8Nvidia RTX 5090Hacker News
Evidence
- Running Qwen3.8 at 255 tok/s on a single 5090 · Ir0n · 13
API: https://socialmediatrends-api.osmike.com/v1/trends/2164021