In the past hour, total liquidations across the network reached $36.32 million, with $3.23 million from long positions, with $33.09 million from short positions. BTC liquidation $14.89 million, ETH liquidation $1.68 million.

According to Coinglass data, $91.5103 million in positions were liquidated across the entire Bitcoin network in the past 24 hours, including $48.4514 million in long positions and $43.0589 million in short positions. Specifically, $3.6984 million in Bitcoin long positions and $7.7529 million in short positions were liquidated, while $3.0752 million in Ethereum long positions and $3.6574 million in short positions were liquidated. A total of 72,107 people were liquidated globally in the past 24 hours. The largest single liquidation order was for the HYPE-USD trading pair on the Hyperliquid platform, valued at $721,500.

According to Coinglass data, if the price of Bitcoin falls below $61,009, the cumulative liquidation intensity of long positions on major centralized exchanges will reach $896 million; if Bitcoin breaks through $66,883, the cumulative liquidation intensity of short positions on major centralized exchanges will reach $756 million.

According to Coinglass data, if ETH falls below $1,597, the cumulative liquidation intensity of long positions on major CEXs will reach $520 million; conversely, if ETH breaks through $1,753, the cumulative liquidation intensity of short positions on major CEXs will reach $444 million.

According to CoinWorld, Anthropic's release of the AI model Claude Fable 5, offering users enhanced reasoning and coding capabilities, could exacerbate security issues facing the crypto market at a time when such problems are troubling. The company launched Claude Fable 5 on Tuesday, the first public model and most powerful version of the Mythos class. The public version boasts enhanced reasoning and coding capabilities while preventing dangerous use. The less restrictive Claude Mythos 5 is only available to vetted cybersecurity users. Experts say Mythos is capable of detecting zero-day vulnerabilities and turning them into attacks. Anthropic states that the software intercepts attack vectors by detecting high-risk requests and triggers a fallback to the weaker Claude Opus 4.8 in less than 5% of sessions. Nevertheless, Anthropic acknowledges that the system cannot be completely secure and anticipates continued attempts by attackers.