Bitget has launched BTC withdrawal services as scheduled.

Beating AI News (Dongcha) – NaiveAI, founded by Tsinghua University associate professor Dai Jifeng, has just released its first model, N0.5-Flash. The company’s core proposition is “AI develops AI”, claiming the model participated in architecture design, code writing, and inference optimization. The official announced peak single-stream inference speed reaches 2122 tokens per second (tok/s). Yang Xinyu, a researcher at Moonshot AI, subsequently publicly questioned this claim. He noted that a speed of 2000 tok/s is achievable, but NaiveAI chose an overly weak baseline for comparison: the company used SGLang’s ~300 tok/s as a reference, while Xiaomi previously ran its larger-parameter MiMo-V2.5-Pro at over 1000 tok/s. A lower baseline makes the resulting improvement appear more dramatic. Yang also rejected NaiveAI’s description of its architectural innovation. N0.5-Flash is modified from Xiaomi’s MiMo-V2.5 Base, replacing the original global attention with DSA. NaiveAI ultimately adopted DSA + GQA. Yang argued that this is more of an engineering trade-off based on an existing base model, which also reduces GPU utilization, and does not qualify as a significant architectural breakthrough. He stated that NaiveAI has proven models can be accelerated via continued training and engineering optimizations, but has not yet demonstrated that “AI developing AI” itself delivers breakthroughs beyond what human teams can achieve.

According to TradingBeats' monitoring, the address starting with 0xc36a on Hyperliquid still holds 1 billion long positions in the PUMP contract, with an average entry price of roughly $0.0036. The current value of this position is approximately $4.99 million, with an unrealized profit of around $1.4 million. PUMP is now trading at about $0.005, some 71% higher than its average entry price. Additionally, the address also holds around 10,800 long positions in the ZEC contract, with an average entry price of roughly $1,206, a position value of approximately $16.66 million, and an unrealized profit of $3.67 million, with no signs of exiting the position so far.

Today, X user @shuangfei8 claimed his MEXC exchange account was hacked, resulting in losses of approximately $340,000, and questioned the platform for failing to revoke the attacker’s API, which led to the full withdrawal of his assets. In response, MEXC officials stated that they have completed an initial investigation and put forward corresponding solutions. The platform will subsequently communicate one-on-one with the user via email to update them on processing progress and required information. MEXC has immediately established a dedicated task force, with specialized staff following up on the relevant account and subsequent handling to safeguard users’ rights and interests.

Bitget CEO Gracy Chen introduced the "Peer Program" during a live stream titled "Bitget Security Incident Review". The program is divided into two parts: For non-VIP users, over the next month, Bitget will return 30% of the net handling fees retained by non-institutional users as exclusive user bonuses, while also rolling out VIP-related benefits. For Pro clients and market makers, an institutional version of the Peer Program will be launched, providing rebates on handling fees and commissions, and extending the Pro level protection period to November 30.

CryptoQuant analyst Darkfost reports that Bitcoin’s Short-Term Holder (STH) Market Value to Realized Value (MVRV) has reached its highest level since November 2025. Momentum accelerated sharply between August and September, mainly driven by the price breaking through the STH cost benchmark of roughly $73,100. Current data shows the STH MVRV stands at 1.15, meaning short-term holders hold an average unrealized profit of around 15% and have returned above the realized value threshold. Darkfost emphasized this level requires close monitoring, as it may prompt short-term holders to start taking profits.

Xie Jiayin, head of Bitget’s Chinese-language region, stated during the "Bitget Incident Review" live stream that attackers exploited a vulnerability in a third-party security product to steal internal network access credentials, forged withdrawal instructions for the wallet system, tricked the wallet into executing abnormal transfers that passed risk checks, and successfully completed the attack. The attackers did not use common viruses or malware; instead, they disguised their actions and covered their tracks by leveraging real identities and routine maintenance operations throughout, making this a relatively sophisticated targeted attack. Following the incident, Bitget has fully ruled out both the possibility of private key leaks and internal misconduct. The hackers exploited the third-party security product’s vulnerability to steal Bitget’s internal credentials and impersonated legitimate identities.