Press Space to continue
DeepSeek V4.1-Flash outperforms Opus 5 at 2.5% of the API cost
Press Space to continue
DeepSeek’s new V4.1-Flash absolutely mogs Opus 5 and GPT-5.6 Sol on DeepSWE v1.1 and a few others. TL;DR: > ~2.5% of Opus 5’s API cost > it uses a new architecture > new pretraining techniques > much larger scale RL post-training https://x.com/Hesamation/status/20979690…
DeepSeek@deepseek_aiSep 10
🚀 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
