DeepSeek-V4.1-Flash
Open 552B MoE model with a new causal encoder-decoder architecture and native vision; DeepSeek phases out V4-Pro.
🚀 Introducing DeepSeek-V4.1-Flash: smarter, faster, more efficient.
Official source
DeepSeek-V4.1-Flash: Smarter, Faster, More Efficientapi-docs.deepseek.com · published
Benchmarks
| Benchmark | Score | Human-verified |
|---|---|---|
| Terminal-Bench 4.0 | 31.2% | Yes |
| DeepSWE v1.1 | 74.2% | Yes |
Reported by DeepSeek
- DeepSeek-V4.1-Flash: Smarter, Faster, More Efficientapi-docs.deepseek.com · 10 Sep 2026
Original image
