BREAKING: DeepSeek released DeepSeek-V4.1-Flash-1017, a new open-weight model
DeepSeek has released DeepSeek-V4.1-Flash-1017, a new open-weight model claimed to reach frontier-level performance, scoring 77% on the DeepSWE 1.1 benchmark 🔥. The "Flash" branding signals a speed-optimized variant balancing performance and inference cost. Competition among open frontier models keeps heating up; real-world performance and the exact open license deserve a closer look.
SOURCES
- Testing Catalog2026-10-17 08:54