Performance Surpasses Opus! Anthropic Leaked Document Reveals: The New Generation Super Model Claude Mythos Is Already in Testing

marsbitОпубліковано о 2026-03-27Востаннє оновлено о 2026-03-27

Анотація

According to leaked internal documents from Anthropic, the company's highly anticipated next-generation AI model, Claude Mythos, is currently in a secret testing phase. The documents reveal a new model tier named "Capybara," which represents a major technological leap, featuring a larger scale and superior intelligence that surpasses the flagship Claude Opus model. The leak also highlights significant concerns within Anthropic regarding unprecedented cybersecurity risks associated with Claude Mythos, prompting a cautious release strategy to balance advanced capabilities with safety. This development is expected to significantly raise the benchmark for large language models, intensifying competition in the AI industry and pushing the evolution of models toward deeper logical reasoning and complex task handling. The official release date for Claude Mythos remains unannounced.

The competition in computing power and intelligence in the field of artificial intelligence is entering a new phase as new models from top-tier labs are exposed.

On March 27, according to media reports citing internal leaked documents from Anthropic, the highly anticipated new generation super model Claude Mythos has now entered a secret testing phase. This leaked blog draft not only showcases the model's powerful performance but also sparks a fresh round of discussion on AI safety.

Defining a New Tier: The Leap from "Opus" to "Capybara"

The leaked document discloses a completely new model tier designation—Capybara. This tier represents the most groundbreaking technological leap in Anthropic's history:

Intelligence Ceiling: The document clearly states that Capybara corresponds to a new, larger-scale tier with a higher level of intelligence.

Surpassing the Flagship: Its comprehensive capabilities have fully surpassed those of the previously industry-benchmark Claude Opus model.

Naming Association: Internal information indicates that Capybara and Mythos most likely refer to different expressions of the same underlying architecture.

The Two Sides of the Coin: Unprecedented Cybersecurity Risks

Alongside the soaring intelligence level, Anthropic internally has also expressed high alert regarding the potential demonstrated by Claude Mythos.

Risk Assessment: The leaked document shows that the company believes this model presents unprecedented cybersecurity risks.

Safety Countermeasures: This risk warning also explains why Anthropic has always maintained a cautious release节奏, attempting to find a stricter balance between pursuing "the most powerful intelligence" and "human safety".

Industry Shockwaves: The Large Model Hierarchy Faces a Reshuffle

As OpenAI's most formidable competitor, Anthropic's move with this new model undoubtedly drops a bombshell on the entire industry:

Competition Escalation: The emergence of Claude Mythos means the baseline for large model capabilities will once again be significantly raised.

Technological Evolution: Judging from the currently disclosed information, the next generation of models is evolving from mere conversational ability towards deeper logical reasoning and complex task handling.

Conclusion: Searching for the "Mythical" Boundaries of AI

Although the official release date for Claude Mythos has not yet been set, the outline of "stronger intelligence" is already clearly visible. As AI's intelligence level begins to surpass the cognitive boundaries of humanity's past, how to harness this power will become a common challenge faced by Anthropic and even global tech giants.

Пов'язані питання

QWhat is the name of Anthropic's new AI model that is currently in secret testing, as revealed by leaked documents?

AClaude Mythos

QAccording to the leak, what is the name of the new model tier that represents a major technological leap for Anthropic and is associated with Claude Mythos?

ACapybara

QWhich existing Anthropic model does the new Claude Mythos model reportedly surpass in overall capability?

AClaude Opus

QWhat major concern does Anthropic have regarding the Claude Mythos model, as mentioned in the leaked documents?

AIt presents unprecedented cybersecurity risks.

QHow is the emergence of Claude Mythos expected to impact the broader AI industry, according to the article?

AIt will significantly raise the benchmark for large model capabilities and force a reshuffling of the industry's top tier.

Пов'язані матеріали

Hands-on with Hunyuan Hy3 Preview: Tencent's AI, Finally Competitive?

Tencent's Hunyuan AI team has released its latest language model, Hy3 preview, marking a significant step forward for the company's AI capabilities. With 295B total parameters and support for 256K context length, the model employs a mixture-of-experts architecture. It shows improvements in complex logic, instruction following, contextual learning, code generation, and agent task execution. In testing, Hy3 preview demonstrated strong performance in multi-step logical reasoning but showed occasional instability in identifying traps in trick questions. It performed well in extracting key information from disordered meeting transcripts and accurately followed new linguistic rules. As an AI agent, it successfully built functional applications like a Snake game and generated data analysis dashboards, though it sometimes fell short in fully completing complex open-ended tasks. In natural language use, it produced coherent and stylistically appropriate narratives with reduced “AI-like” tone. Priced competitively, Hy3 preview is already integrated into Tencent’s key products, including Tencent Cloud and WorkBuddy. While not leading in every benchmark, it represents a solid, practical model that signals Tencent’s renewed momentum in AI development.

marsbit1 год тому

Hands-on with Hunyuan Hy3 Preview: Tencent's AI, Finally Competitive?

marsbit1 год тому

From Theft to Re-entry: How Was $292 Million "Laundered"?

A sophisticated crypto laundering operation was executed following the $292 million hack of Kelp DAO on April 18. The attack, attributed to the North Korean Lazarus group, began with anonymous infrastructure preparation using Tornado Cash to fund wallets untraceably. The hacker exploited a vulnerability in Kelp’s cross-chain bridge, stealing 116,500 rsETH. To avoid crashing the market, the attacker used Aave and Compound as laundering tools—depositing the stolen rsETH as collateral to borrow $190 million in clean, liquid ETH. This move triggered a bank run on Aave, causing an $8 billion drop in TVL. After consolidating funds, the attacker fragmented them across hundreds of wallets to evade detection. A major breakpoint was THORChain, where over $460 million in volume—30 times its usual activity—was processed in 24 hours, converting ETH into Bitcoin. This shift to Bitcoin’s UTXO model exponentially increased tracing complexity by shattering funds into countless untraceable fragments. The final destination was Tron-based USDT, the primary channel for illicit crypto flows. From there, funds were cashed out via OTC brokers in China and Southeast Asia, using unlicensed underground banks and UnionPay networks outside Western sanctions scope. Ultimately, the laundered money supports North Korea’s weapons programs, which rely heavily on crypto hacking for foreign currency. The incident underscores structural challenges in DeFi: its openness, composability, and lack of central control make such laundering not just possible, but inherently difficult to prevent.

marsbit1 год тому

From Theft to Re-entry: How Was $292 Million "Laundered"?

marsbit1 год тому

Behind DeepSeek V4's Stunning Debut: Silicon Valley Is 'Building Walls,' China Is 'Paving Roads'

China's AI landscape is witnessing a strategic divergence from Silicon Valley’s closed-source competition to a collaborative open-source ecosystem. On April 24, DeepSeek released V4, a top-ranked open-source model on Hugging Face, featuring breakthroughs like million-token context length with minimal KV cache and native support for domestic chips like Huawei’s Ascend. Similarly, Kimi’s K2.6, released days earlier, also adopted open-source principles. Unlike U.S. giants such as OpenAI and Anthropic—locked in revenue disputes and tactical product clashes—Chinese firms embrace shared innovation. DeepSeek and Kimi openly build on each other’s advances, like the MLA architecture and Muon optimizer, avoiding redundant R&D and driving down costs. DeepSeek V4 focused on pushing base model capabilities, while Kimi specialized in Agent-based applications. Although U.S. firms lead in revenue and valuation, China’s open-source models achieve comparable performance at a fraction of the cost (e.g., DeepSeek V3 trained for $5.58M vs. GPT-5’s $500M+). With token usage growing exponentially, China’s collaborative model promises scalable, affordable AI built on domestic hardware, shaping a more accessible path to AGI.

marsbit1 год тому

Behind DeepSeek V4's Stunning Debut: Silicon Valley Is 'Building Walls,' China Is 'Paving Roads'

marsbit1 год тому

OpenClaw vs. Hermes: Which One Is Right for You?

Two leading AI agent projects, OpenClaw and Hermes, represent a fundamental divergence in design philosophy rather than just a feature competition. OpenClaw, a GitHub phenomenon, prioritizes user control and customization, offering a powerful, flexible platform for building and orchestrating agents. It boasts a massive ecosystem with over 44,000 skills, extensive model flexibility (supporting Anthropic, OpenAI, and others), and deep integrations across multiple messaging platforms. In contrast, Hermes Agent, developed by Nous Research, champions automation and efficiency. It is designed for lower cost and a lower barrier to entry, learning and improving automatically from user workflows. It runs tasks in isolated environments for security and offers significantly lower token costs out-of-the-box. The choice between them mirrors historical tech divides like Windows vs. Mac. OpenClaw is for users who want maximum control, customization, and don't mind a complex setup. Hermes is the smarter default for those seeking an affordable, easy-to-use agent that learns autonomously. The real competition is not about which is better, but which philosophy—a programmable personal OS or a self-evolving work proxy—better suits the user's needs for control, cost, and ease of use.

marsbit1 год тому

OpenClaw vs. Hermes: Which One Is Right for You?

marsbit1 год тому

Google and Amazon Simultaneously Invest Heavily in a Competitor: The Most Absurd Business Logic of the AI Era Is Becoming Reality

In a span of four days, Amazon announced an additional $25 billion investment, and Google pledged up to $40 billion—both direct competitors pouring over $65 billion into the same AI startup, Anthropic. Rather than a typical venture capital move, this signals the latest escalation in the cloud wars. The core of the deal is not equity but compute pre-orders: Anthropic must spend the majority of these funds on AWS and Google Cloud services and chips, effectively locking in massive future compute consumption. This reflects a shift in cloud market dynamics—enterprises now choose cloud providers based on which hosts the best AI models, not just price or stability. With OpenAI deeply tied to Microsoft, Anthropic’s Claude has become the only viable strategic asset for Google and Amazon to remain competitive. Anthropic’s annualized revenue has surged to $30 billion, and it is expanding into verticals like biotech, positioning itself as a cross-industry AI infrastructure layer. However, this funding comes with constraints: Anthropic’s independence is challenged as it balances two rival investors, its safety-first narrative faces pressure from regulatory scrutiny, and its path to IPO introduces new financial pressures. Globally, this accelerates a "tri-polar" closed-loop structure in AI infrastructure, with Microsoft-OpenAI, Google-Anthropic, and Amazon-Anthropic forming exclusive model-cloud alliances. In contrast, China’s landscape differs—investments like Alibaba and Tencent backing open-source model firm DeepSeek reflect a more decoupled approach, though closed-source models from major cloud providers still dominate. The $65 billion bet is ultimately about securing a seat at the table in an AI-defined future—where missing the model layer means losing the cloud war.

marsbit7 год тому

Google and Amazon Simultaneously Invest Heavily in a Competitor: The Most Absurd Business Logic of the AI Era Is Becoming Reality

marsbit7 год тому

Торгівля

Спот

Ф'ючерси