The Chinese labs did it again
Open weights, ridiculous API pricing, and a Flash model that beats their previous flagship on several real agent benchmarks
When can we test it on Hermes?
@NousResearch
🚀 Introducing DeepSeek-V4.1-Flash: smarter, faster, more efficient.
🔹 Introducing the smallest model in our new architecture family, with native visual understanding.
🔹 Designed for greater capability, faster inference, higher throughput, and scaling to larger models.
1/6
もっと見る