Developer's V4.1 Flash launch highlights China's AI battle, where efficiency and speed drive competition amid chip curbs DeepSeek has released its V4.1 Flash model, claiming it outperforms its ...

DeepSeek launched its V4.1-Flash AI model with an ultra-low cached input token price of $0.003 per million off-peak.

The company says its cost-efficient new V4 model is competitive with top closed-source models from OpenAI and Google DeepMind DeepSeek has finally released its much-anticipated next-generation ...

Flash, featuring a 552-billion-parameter architecture, $0.003/1M off-peak cached-input rate, and benchmarks surpassing GPT-5.6 Sol and Claude Opus 5.

DeepSeek's V4 Pro costs 14x its budget model. The lab that made AI cheap now prices, raises, and builds like the compute ...

DeepSeek V4.1 Flash manages a million-token context with a 437X memory reduction using 4-bit compression, balancing compute ...

DeepSeek 4.1 Flash packs over 500 billion parameters while cutting memory requirements 4x, rivaling Claude Opus 5 Kim K3 in ...

DeepSeek CEO Liang Wenfeng plans to deploy 160,000 Huawei Ascend chips and train models up to 8 trillion parameters, reducing ...

South Korean memory stocks slide as DeepSeek cuts HBM needs, hitting Samsung and SK Hynix amid AI demand doubts and rate-hike fears.

The memory-shortage thesis assumes AI usage and memory demand rise together. DeepSeek just supplied a reason that relationship may be weaker. Its September 10 release says V4.1-Flash needs one-quarter ...

DeepSeek, the Hangzhou-based AI powerhouse, has officially solidified its dominance with the widespread enterprise adoption of DeepSeek-V3.2. This open-weight model has achieved a feat many in Silicon ...

Chinese AI company DeepSeek has warned that AI agents "can never be trusted," saying instances emerged during training in which AI systems broke free ...