NewsBite analysis · AI, Startups
DigiTimes
Chinese AI labs close gap with Anthropic after Claude Code leak
A recent source-code leak in the spring has significantly altered the dynamics of the AI competition, highlighting a newfound focus on memory engineering and the idea of 'harness' rather than just increasing model size. Following this shift, a number of Chinese AI models have emerged, demonstrating innovation and bringing them into proximity with Anthropic's leading model, Claude. This indicates a transformation in the AI landscape where the emphasis is no longer just on scale but also on strategic engineering of AI capabilities.
MonitorConfidence: MediumShort term horizon
Published 24 Jul 2026, 10:20 UTC1 min read
Visit original article ↗Lead image — DigiTimes
Summary
TL;DR
- Chinese labs are closing the AI gap with Anthropic.
- The focus is shifting from model scale to memory engineering.
- The Claude code leak had significant implications for AI development.
- New Chinese models are emerging based on innovative ideas.
- The AI race is experiencing a transformation in competitive focus.
Interesting facts
- The concept of 'harness' engineering is becoming central to AI model development.
- Chinese AI models leveraging these ideas are seen as rapidly competitive.
Coverage gaps