DeepSeek v4.1 performance claims on OpenRouter: 172 tokens per second, 94% cache rate
A provider, citing OpenRouter figures, says it serves 30% of DeepSeek v4.1 requests on the platform and has the fastest throughput among supported serverless endpoints.
TLDR
In a September 16, 2026 post citing OpenRouter, a provider says it serves 30% of the platform’s DeepSeek v4.1 requests at 172 tokens per second, compared with 107 for the next-fastest endpoint. It also claims the best cache rate, at 94%, and top-decile latency among all supported serverless endpoints.
Combined views
3.2K
1 Source, first seen 14d ago