DeepSeek V4.1 Flash hits 532 tokens/s on Inco, fastest output on Artificial Analysis

songhan_mit · x · 2026-09-18

Inference provider Inco announced DeepSeek V4.1 Flash is live with 532 tokens/s output, ranking #1 output speed on Artificial Analysis with a clear lead.

Blogger songhanmit amplified the point: what matters isn't just being fast but iterating fast — high throughput accelerates the entire loop of interactive development and experimentation. The service is live with benchmark links attached.

Related event: DeepSeek Launches V4.1-Flash with Native Vision, Tops Speed Charts(3 posts)→

Original post →

More from Infra

Infra channel →