The Metaverse Upgrades Again: Baidu World 2023 Introduces Multiple New AI Features


Market rumors claimed AI unicorn Moonshot AI plans a Hong Kong IPO this month to raise $3 billion, which the company has denied. However, it recently made frequent capital and organizational moves, having just completed an unprecedented Series F financing last month.....
On July 31, the official DeepSeek-V4-Flash API launched. Its agent performance rivals the Pro version, breaking the traditional stratification of weak vs. strong models. With 284B total parameters, it achieved top scores on benchmarks like Terminal Bench 2.1, marking a key step for Chinese LLMs in the agent domain.....
Tencent open-sources the unified training framework AngelSpec, which reduces the cost of autoregressive decoding for large models and improves inference throughput through collaborative schemes and heterogeneous workload adaptation. It adopts differentiated training strategies for different text characteristics in scenarios, such as using a lightweight and stable approach for high-entropy multi-turn dialogues, achieving efficient inference in real-world scenarios.
Alibaba Cloud has rebranded the "Alibaba Cloud Developer" WeChat Official Account to "Qwen AI Platform", positioning it as a platform designed for Agents, driving AI productivity. This move aligns with the trend of AI evolving towards intelligent agents, focusing on cutting-edge technologies in large models and Agents. Relying on the Qwen model and Alibaba Cloud technology, it will launch an all-around upgrade of AI content and developer services.
Sina's VibeThinker-3B, a 3B-parameter model, rivals much larger models in math and coding benchmarks, surpassing top products on certain competition tasks. Its success stems from unique training on Alibaba's Qwen2, challenging the 'bigger is better' paradigm.....