DeepSeek says new Flash AI model beats Kimi K3 on cyber, coding benchmarks
SCMP Techen

DeepSeek has released its V4.1 Flash model, claiming it outperforms its previous flagship while cutting inference costs and boosting speeds – the latest salvo in China’s aggressive price-and-performance war. The Chinese artificial intelligence developer said on Thursday that V4.1 Flash used a new “Causal-Encoder-Decoder” architecture. While built on a massive 552 billion-parameter framework, it relies on a Mixture-of-Experts (MoE) design. In traditional AI, every query runs through the entire...
This is a short summary published by AI Global Wire. The full article is owned and hosted by SCMP Tech — open it there to read it in full.
Read the full story at SCMP Tech- DeepSeek
- Verktyg
Related AI news
- 中國民眾好難,iPhone Duo 中版沒 AI、海外版沒門號TechNews (TW) · September 10, 2026
- Clearview AI Is Testing an AI Tool That Would Let Cops Unearth Your Life OnlineWIRED AI · September 10, 2026
- DeepSeek debuts DeepSeek-V4.1-Flash, its smallest model built on a new Causal Encoder-Decoder architecture, with 552B backbone parameters and 1M-token context (Reuters)Techmeme · September 10, 2026
- DeepSeek V4.1-Flash targets lower AI agent costsTech in Asia · September 10, 2026
- Top AI spenders cut per-employee costs by nearly 10 percent in AugustThe Decoder · September 10, 2026
- Bybit rolls out AI tool for trading, account managementTech in Asia · September 10, 2026