PANews 3月27日消息，据《财富》报道，Anthropic 承认正在测试一款较现有产品有“跃迁式提升”的新模型 Claude Mythos，并已向小范围早期客户开放。泄露的内部草稿显示，该模型内部代号为“Capybara”，性能全面超过此前最强的 Claude Opus 4.6，在软件编程、学术推理和网络安全测试中得分大幅提升。Anthropic 在未公开的数据缓存中留下近 3000 份未发布资产，包括介绍 Claude Mythos/Capybara 的博文草稿及一场欧洲 CEO 邀请制闭门峰会的资料。公司称该模型在网络攻防方面“远超其他 AI”，计划先向网络安全防御方定向投放，并因其运行成本高、风险大而暂不全面开放。
Anthropic泄露文档显示正测试新一代AI模型Claude Mythos
Favorite
Share
Disclaimer: This article is copyrighted by the original author and does not represent MyToken’s views and positions. If you have any questions regarding content or copyright, please contact us.(www.mytokencap.com)contact
About MyToken:https://www.mytokencap.com/aboutusArticle Link:https://www.mytokencap.com/news/569950.html
Previous:比特币矿企加速告别挖矿时代，MARA大举卖币砸向AI
Next:逆势增长：MCD穿越周期的价值密码
Related Reading
manadia and PvPfun Partner to Make AI-Built Onchain Apps Verifiable And Trustworthy
manadia partners with PvPfun to add verifiable data settlement to AI-powered onchain apps, making us...
blockchainreporter2026-03-27 14:00:00
Why Is Bitcoin Crashing Today? $14B Options Expiry, Iran War Escalations, & More
The post Why Is Bitcoin Crashing Today? $14B Options Expiry, Iran War Escalations, & More appeared f...
CoinPedia2026-03-27 13:42:05
UK Slaps Sanctions On $20B Crypto Black Market Tied To Southeast Asia Scam Rings
Blockchain analytics firm Chainalysis puts the number at close to $20 billion — the estimated volume...
NewsBTC2026-03-27 13:30:03