DIGITIMES

DeepSeek V4 introduces utility-style AI pricing in shift beyond China's LLM price war

DeepSeek will launch the official version of its V4 large language model (LLM) in mid-July alongside peak and off-peak API ...

DeepSeek open sources DSpark, a new framework to speed up LLM inference by up to 85%

DSpark can make decoding faster, but acceptance quality still determines how much speed the system actually realizes.
NewsBytes

This is China's biggest AI model trained on local chips

China achieves a breakthrough in AI development as Meituan unveils LongCat-2.0, the nation's largest model trained entirely ...

India Risks Permanent Dependence On Foreign AI Tools Unless It Builds Its Own 'DeepSeek-Like' Sovereign Model: Bernstein

India risks becoming permanently dependent on foreign AI models unless it develops sovereign large language models, warns ...

India needs its own DeepSeek to avoid AI dependence, warns Bernstein

India must build its own sovereign large language models or risk becoming permanently dependent on foreign AI, according to a ...
Tech Times

Most AI Models Would Run Your Company Into the Ground, Princeton’s CEO-Bench Finds

Princeton’s CEO-Bench gave 14 AI models $1 million to run a simulated SaaS startup for 500 days. Most went bankrupt or lost ...
techtimes

DeepSeek Releases DSpark: Speculative Decoding Makes V4 Up to 85 Percent Faster

In this photo illustration, the DeepSeek app is displayed on an iPhone screen on January 27, 2025 in San Anselmo, California. Newly launched Chinese AI app DeepSeek has surged to number one in Apple's ...

VCs Pour $3 Billion Into AI's Next Big Bet: World Models

Venture investors poured more than $3 billion into world model startups in 2026, betting AI that can simulate the physical ...

Xiaomi's HarnessX rewrites its own AI scaffolding mid-task — and smaller models gain the most

Xiaomi's HarnessX autonomously rewrites AI agent harnesses mid-execution, delivering +14.5% avg performance gains — and +44% ...

Small Language Models Outperform Frontier AI On Cost, Speed And Accuracy

Bigger has defined AI from day one. New data says task-specific small models beat frontier LLMs on accuracy, cost and speed — ...
TRT World

New AI model by China's Meituan trained on domestic chips, company says

Beijing has accelerated efforts to develop advanced domestic chips and reduce its reliance on American hardware amid an ...

IBM’s 1-nanometer chip, Google’s AI policy pitch, and Chinese AI

The biggest AI developments, decoded. 1 July, 2026.