Solana launches public beta test of its Alpenglow upgrade, aiming to reduce transaction finality to 150 milliseconds.

Per Bitget market data, Hong Kong-listed AI application stocks extended their afternoon decline: Zhipu fell over 10%, Kingdee International dropped more than 7%, while Meitu, MINIMAX-W and other related stocks also tracked lower.

Whales keep buying $UNI today! A newly created wallet, 0xEFC4, bought 269,477 $UNI ($2.84M). 0xf415 bought another 138,442 $UNI ($1.23M).

Whale 0xA799 bought 788,000 $UNI ($4.93M) at $6.26 6 days ago and sold at $8.85, making $2.04M in less than a week!

According to official announcements, NEAR has been deployed on Hyperliquid's spot market. Users can trade the NEAR/USDC pair on Hyperliquid. The official added that per Hyperliquid's procedures, NEAR will take several more days to appear on the platform's Strict List.

According to an official announcement, Binance will launch USDT spot trading pairs for Axe Compute (AGPUB), AMC Entertainment (AMCB), and Cypherpunk Technologies (CYPHB) at 20:00 China Standard Time (UTC+8), and simultaneously introduce spot algorithmic trading bots. The relevant trading pairs will enjoy zero maker fees until 07:59 on October 1, while withdrawals will open at 21:00 the same day. The aforementioned tokenized securities are not available to US persons.

Beating AI Insight News Flash. Leiphone reports that over the past six months, multiple domestic large language model (LLM) vendors have begun redoing their pre-training processes, with all issues traced to data quality. One vendor spent nearly a year training a trillion-parameter model, only to have its performance outperformed by a small model with just around 1% of the parameter count. The team ultimately found the root cause lay in the training data: web spam, duplicate corpora, and low-quality annotations had been mixed into the training set, rendering even the largest models ineffective. The case comes from an anonymous source, and the specific company has not been disclosed. Tencent Hunyuan has publicly announced a full redo of its pre-training work. Since February this year, the team has rebuilt its pre-training and reinforcement learning infrastructure. Previous media reports revealed that the old version of Hunyuan had issues including ranking manipulation data being mixed into the training set and inconsistent annotation rules. The new team redefined data standards, cleaned up existing corpora, and Hy3 listed data quality and diversity as key improvement areas. Alibaba and Baidu are also stepping up data governance efforts. Alibaba’s Qwen3 uses the Qwen2.5 series models to clean documents, and has significantly supplemented synthetic math and code data. Baidu’s Ernie 4.5 has added deduplication, low-quality filtering, data mapping, and manual review processes. Neither company has publicly stated that they had to redo work due to dirty data, but both have invested more efforts in data processing for their new-generation models.