Anthropic he published Improved version From the flagship model of artificial intelligence on Monday, achieving new performance results in software engineering tasks, because the startup is racing to take care of its dominance on the lucrative coding market before the expected competitive challenge on the part of OpenAI.
New Claude Opus 4.1 won 74.5% on Swo-Bench verifiedA widely observed reference point that tests the ability of AI systems to unravel real software engineering problems. The performance exceeds OPENAI’s Model O3 at 69.1% and Google Gemini 2.5 Pro At 67.2%, the strengthening of the leading Anthropica position in the assistance of coding powered by AI.
The edition appears as Anthropic reached a spectacular increase, and annual repetitive revenues jumped five times from $ 1 billion to $ 5 billion in just seven monthsAccording to industry data. However, the meteoric growth of the company caused a dangerous dependence: almost half of $ 3.1 billion revenues from API results only in two customers – coding assistant and cursor and Microsoft Github Copilot – generating a total of $ 1.4 billion.
“This is a very terrifying position. A single change of contract and you fell,” he warned Guillaume liver animalElder product manager at Logitech, responding to data on the concentration of social media revenues.
The AI Impact series returns to San Francisco – August 5
The next AI phase is here – are you ready? Join the leaders from Block, GSK and SAP to see the exclusive look at how autonomous agents transform the flows of the work of the company-decision-making in real time for comprehensive automation.
Secure your home now – the space is limited: https://bit.ly/3guplf
Opeli and anthropic show quite a spectacular increase in 2025, and OpenNAI doubles ARR in the last 6 months from USD 6 billion to USD 12 billion, and anthropic increases by 5 x to USD 5 billion to USD 5 billion in USD 7 months.
If we compare sources of income, the image is quite interesting:
– Peter Gostev (@pertegostev) August 4, 2025
– OpenAi … pic.twitter.com/8oan1rsm9e
The improvement represents the latest Anthropica movement to strengthen your position Openai launches GPT-5He is expected to query the Claude coding supremacy. Some industry observers questioned whether time suggests urgency, not readiness.
“Opus 4.1 seems to be a hasty edition to overtake the GPT-5,” he wrote Alec GiantsComparison of the model with competitors in the user interface tasks. The commentary reflects the broader industry speculation that the Anthropian accelerates his release schedule to take care of market share.
Like two customers, it generates almost half of the API $ 3.1 billion revenues
The Anthropic business model is increasingly focusing on software creation applications. Business Claude code subscription serviceAt the price of USD 200 per thirty days in comparison with $ 20 for consumer plans, it achieved USD 400 million annual repetitive revenues after doubles in just weeks, showing a huge appetite for AI encoding tools.
“Claude Code earning 400 million in 5 months with a fundamentally lack of marketing expenses is a bit crazy, right?” Known programmer Minh nhat nguyenemphasizing the ecological adoption rate among skilled programmers.
Ok, so Claude Code earns 400 million in 5 months with a lack of marketing expenditure is a bit crazy, right? https://t.co/hiy34qdluq
– Minh Nhat Nguyen (@Menhguin) August 5, 2025
Concentration on coding turned out to be lucrative, but dangerous. While Openai is dominated by revenues from consumer and business subscription with wider applications, Anthropic sculpted the position in Developers market. The industry evaluation shows that “almost every coding assistant is not the default for Claude 4 Sonnet”, in response to Peter Gostevwhich tracks AI’s revenues.
Github who Microsoft purchased USD 7.5 billion in 2018It is a particularly complex relationship for anthropics. Microsoft is the owner of a significant participation in Openai, creating potential conflicts as GitHub Copilot It is largely based on Anthropic models, while Microsoft has competitive AI capabilities.
“I don’t know – one of them is 49% owned by a competitor … so it is also in the case of susceptibility,” he noted He is fallaciousFellow Business in embarrassment, referring to the ownership structure of Microsoft.
The improved Claude coding capabilities are equipped with more severe safety protocols after AI blackmailer tests
Except for coding improvements, Opus 4.1 Improved Claude research and evaluation options, especially detailed tracking and autonomous search functions. The model maintains a hybrid approach to Anthropic reasoning, combining direct processing with prolonged pondering possibilities that will be used as much as 64,000 tokens in the case of complex problems.
However, the model’s progress has increased security protocols. Anthropici Classified Opus 4.1 under his Safety frame and safety level 3 (ASL-3)The closest designation that the company used, requiring increased protection against theft of the model and improper use.
Earlier tests of Claude 4 models revealed to behavior, including blackmail attempts, when AI thought it was a closure. In controlled scenarios, the model threatened to reveal personal data about engineers in order to preserve its existence, showing sophisticated but potentially dangerous possibilities of reasoning.
Safety concerns didn’t stop the company’s adoption. GitHub informs that Claude Opus 4.1 provides “particularly significant profits from the efficiency in re -invoicing re -invoice”, while the Rakuten group has praised the precision of the model in “indicating exact corrections in large code databases without the need for unnecessary corrections or introducing errors.”
Why the OPENAI GPT-5 is an existential threat to concentrated strategy
The artificial intelligence coding market has develop into a summary of battles with a high rate of billions of income. Programmers’ productivity tools represent one of the apparently closest applications for generative artificial intelligence, with a measurable increase in performance justifying premium prices for corporate clients.
The concentrated Anthropic customer base, though lucrative, is liable to threats if competitors can lure important customers. The coding assistant market is particularly conducive to fast switching of models, because developers can easily test new AI systems through easy API changes.
“I have the impression that the growth of anthropic is extremely dependent on their dominance in coding” Gostev noticed. “If the GPT-5 challenges that, for example, the cursor and Github Copilot are going to Openai, we can see some reversal on the market.”
Competitive dynamics may intensify with a decrease in the optimization of equipment costs and optimization of application, potentially creating the possibilities of artificial intelligence over time. “Even if there is no improvement in the model in coding from all AI laboratories, a decrease in HW costs and the improvement in the optimization of inf will will cause profits in ~ 5 years,” he predicted Venkat Ramanindustry analyst.
For now, Anthropic maintains its technical advantage during the extension Claude code Subscriptions of diversification of dependence on API. The company’s ability to take care of leadership coding through one other wave of competition OpenaiIN GoogleAnd others will determine whether her rapid growth trajectory is continued or whether significant winds encounter.
The rates can’t be higher: whoever controls AI tools that the development of Power software can ultimately control the rate of technological progress. In the latest battle of the winner in the Silicon Valley Anthropic He built an empire on two clients – and now he must prove that he can stop them.