DeepSeek
Tracking DeepSeek's model iteration and open-source strategy, we look at the actual progress of the V3 and V4 series in long-context scenarios, agent capabilities, and domestic computing power adaptation.
Tracking DeepSeek's model iteration and open-source strategy, we look at the actual progress of the V3 and V4 series in long-context scenarios, agent capabilities, and domestic computing power adaptation.
MAI-Base-2 has finished pre-training. On October 15, 2026, Microsoft AI published the details in its official post "Frontier Pre-training at Microsoft...
DeepSeek published a service restoration and compensation notice on its open platform on the evening of October 15, 2026. Between 08:11 and 13:04 Beij...
ARTEX is an open-source agentic penetration-testing tool. On October 7, 2026, US cybersecurity firm CrowdStrike published an investigation reporting t...
The Oct 7 AI briefing rounds up the past 24 hours: money and models are both accelerating. DeepSeek's latest funding round is reported to reach at lea...
DeepSeek V4.1 Flash's newest results were published by ARC Prize on October 6, 2026: in the ARC-AGI (Verified) evaluation, the independent evaluator s...
How far Chinese models go in dodging sensitive questions now has a systematic measurement: the German company Aleph Alpha tested them against 967 topi...
On September 30, 2026, DeepSeek open-sourced infrastructure components for Ascend. DeepSeek announced today via its official WeChat account that it is...
DeepSeek has finally put "making money" on the table. On September 24, The Information reported that DeepSeek's annualized revenue run rate has reache...