Yhn HealthFitness first seen 1 d 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 language model at 255 tokens per second on a single Nvidia RTX 5090 graphics card, without multi-GPU hardware. The result is drawing attention from people tracking how far consumer-grade GPUs can now go for local AI inference, since that speed is fast enough for responsive real-time use on a desktop setup.
Why now: It shows a new consumer GPU delivering server-class inference speeds for a popular open model.
Rank over time, top of the chart is #1. 20 snapshots from 1 d ago to 1 h ago.
Evidence
- Running Qwen3.8 at 255 tok/s on a single 5090 · Ir0n · 14
API: https://socialmediatrends-api.osmike.com/v1/trends/2164021