Humanity’s new Claude 4.1 dominates coding tests a few days before GPT-5 arrives

    Crypto AI Updates012 mins

    Need smarter insights in your inbox? Sign up for our weekly newsletter to get only the things that matter to enterprise AI, data and security leaders. Subscribe now

    Anthropic released an upgraded version of its flagship artificial intelligence model on Monday, achieving new high performance in software engineering tasks to help AI startups maintain dominance in the lucrative coding market ahead of the expected competitive challenges from Openai.

    The new Claude Opus 4.1 model won 74.5% on SWE bench verification. This is a widely viewed benchmark that tests the ability of AI systems to solve real software engineering problems. This performance outperforms Openai’s O3 model at 69.1%, while Google’s Gemini 2.5 Pro surpasses 67.2%, solidifying the main position of humanity in AI-powered coding support.

    Industry data shows that annual recurring revenues have increased five times from five times to $5 billion in just seven months, as annual recurring revenues have achieved spectacular growth. However, the company’s meteor rise created a dangerous dependency. Almost half of the $3.1 billion in API revenue is generated from just two customers, from assistant assistant cursors and coding Microsoft’s Github Copilot.

    “This is a very scary position. A single contract has been changed and you are below,” said Guillaume Leverdier, senior product manager at Logitech, who responded to revenue-focused data on social media.

    The AI Impact Series returns to San Francisco – August 5th

    The next phase of AI is here – Are you ready? Join Block, GSK and SAP leaders to see exclusively how autonomous agents are reshaping their enterprise workflows, from real-time decision-making to end-to-end automation.

    Secure your spot now – Space is limited: https://bit.ly/3guplf

    This upgrade represents the latest moves of humanity to strengthen its position before Openai launches GPT-5, and is expected to challenge Claude’s coding advantage. Some industry watchers questioned whether timing suggests more urgency than preparation.

    “Opus 4.1 feels like a rushed release to get ahead of the GPT-5,” wrote Alec Velikanov, comparing the model to its user interface task competitors. The comments reflect broader industry speculations that humanity is accelerating its release schedule to maintain market share.

    How two customers generate almost half of humanity’s $3.1 billion API revenue

    Anthropic’s business model is increasingly focused on software development applications. The company’s Claude code subscription service costs $200 a month compared to the $20 consumer plan, doubled in just a few weeks, and then reaches $400 million in annual recurring revenues, indicating the vast company’s desire for AI coding tools.

    “Claude Code makes 400 million code in five months, and basically doesn’t have marketing spending. That’s a bit weird, isn’t it?” Famous developer Minh Nhat Nguyen highlights the organic adoption rates of professional programmers.

    The focus of coding has proven advantageous but dangerous. Openai controls consumer and business subscription revenues with a wider range of applications, while Anthropic carries the position of commander in the developer market. According to Peter Gostev, who tracks AI companies’ revenue, industry analysis shows that “almost all coding assistants are defaulted to Claude 4 Sonnet.”

    Github, which Microsoft acquired for $7.5 billion in 2018, represents a particularly complex relationship between humanity. Microsoft owns a significant stake in Openai, and Github Copilot relies heavily on the human model, creating potential conflicts.

    “I’m Dunno. One of them is 49% owned by a competitor, so there are vulnerabilities too,” said Siya Mali, a business fellow who references Microsoft’s ownership structure.

    Claude’s enhanced coding capabilities come with more stringent safety protocols after AI Blackmail testing

    Beyond coding improvements, OPUS 4.1 has enhanced Claude’s research and data analysis capabilities, particularly in detail with tracking and autonomous search capabilities. This model maintains human hybrid inference approaches, combining direct processing with extended thinking capabilities that allow for up to 64,000 tokens to complex problems.

    However, advances in the model are accompanied by increased safety protocols. Under the AI Safety Level 3 (ASL-3) framework, the Human Classification OPUS 4.1 is the strictest designation applied by the company, requiring greater protection against theft and misuse of models.

    Previous tests of the Claude 4 model revealed behavioral behavior, including a fearful email attempt when AI believed it was facing a shutdown. In a controlled scenario, the model threatened to reveal personal information about the engineer and maintain its presence, demonstrating sophisticated but potentially dangerous reasoning capabilities.

    Safety concerns have not thwarted companies’ recruitment. Github reports that Claude Opus 4.1 provides “particularly significant performance improvements in multi-file code refactoring,” while Rakuten Group praised the model’s accuracy of “identifying accurate fixes within a large codebase without introducing unnecessary tweaks or bugs.”

    Why Openai’s GPT-5 poses an existential threat to a strategy focused on human developers

    The AI coding market has become a battlefield for high stakes, worth billions of revenue. Developer productivity tools are some of the clearest immediate applications of AI that can be generated, and measurable productivity justifies premium pricing for corporate customers.

    Humanity’s concentrated customer base is advantageous, but creates vulnerability when competitors can seduce key clients. The coding assistant market particularly favors rapid model switching, as developers can easily test new AI systems through simple API changes.

    “My sense is that human growth is highly dependent on their control in coding,” Gostev pointed out. “If the GPT-5 is trying that, for example, if the cursor and Github co-pilot switch to Openai, you might see some reversal in the market.”

    As hardware costs decrease, inference optimization improves, and potentially commoditizing AI capabilities over time, competitive dynamics can intensify. “Even without model improvements for coding from all AI labs, lower HW costs and improvements solely in INF optimization will benefit in five years,” predicted industry analyst Venkat Raman.

    For now, humanity is expanding its Claude code subscriptions and diversifying beyond API dependencies, while still maintaining its technological edge. Through the next wave of competition with Openai, Google and others, the company’s ability to maintain coding leadership determines whether its rapid growth trajectory continues or faces significant headwinds.

    Stakes will not be expensive. Anyone who controls AI tools that power software development can ultimately control the pace of technological advancements themselves. In all the battles of Silicon Valley’s latest winners, humanity built an empire with two clients.



    Source link

    Related News