Anthropic unveils faster, cheaper Opus 4.5

12/25, AI Models & Tools, Global

04/12/2025

Anthropic has launched Opus 4.5, its latest flagship model, improving coding performance, efficiency, and the overall user experience. The model fixes previous conversation cut-offs in Claude by summarizing key points when the context limit is reached, preventing abrupt endings. Opus 4.5 achieves over 80% accuracy in relevant benchmarks, outperforming GPT-5.1-Codex-Max and Gemini 3 Pro in coding tasks, while being less vulnerable to prompt-injection attacks.

It’s also far more token-efficient, achieving top results with significantly fewer tokens. Other updates include a new “Effort” parameter for developers, Claude Code in desktop apps, and significantly lower API pricing.

Read the article

Updates from the world of artificial intelligence

Recent News

13. Januar 2026

Use of AI in school deemed cheating

The German Administrative Court in Hamburg ruled that the use of AI tools such as ChatGPT to write homework assignments can be considered cheating, even if the school has not expressly prohibited it.

13. Januar 2026

New attack exposes AI security vulnerabilities

Researchers have discovered a new attack on ChatGPT that exploits vulnerabilities in the AI's processing of input prompts. Attackers can skim sensitive user data and embed permanent instructions in the model's memory.

13. Januar 2026

Bosch invests 2.9 billion in AI

Bosch plans to invest around 2.9 billion € in artificial intelligence by 2027 and make AI an integral part of its manufacturing and supply chains, moving beyond pilot projects. The company uses AI to evaluate sensor and camera data and identify quality issues at an early stage.

13. Januar 2026

Protective measures against potential AI turbulence

The bursting of a potential AI-driven stock market bubble could have significant consequences for personal finances. The rapidly rising valuations of AI companies may be excessive, and a sharp fall in share prices would affect the markets as a whole.

13. Januar 2026

How AI is changing jobs and careers

In 2026, artificial intelligence will continue to transform the job market and influence how people work, are hired and plan their careers. Companies are increasingly using AI in recruitment processes and everyday work.

13. Januar 2026

AI interacts with refrigerator

At CES 2026, Samsung unveiled a smart refrigerator that uses Google Gemini and internal cameras to recognise what is inside the refrigerator. It offers assistance with creating shopping lists, meal planning and cooking.

13. Januar 2026

AI PCs will augment, not replace humans

At CES 2026, Lenovo CEO Yuanqing Yang stated that by the end of 2026, around half of all business-grade PCs will be AI-enabled. He emphasised that AI is unavoidable and has a profound impact, but will not replace humans.

8. Januar 2026

AI is changing shopping behaviour and brand choice

Artificial intelligence is increasingly changing the shopping experience. More and more consumers are using generative AI tools to make purchasing decisions rather than just finding products.

8. Januar 2026

Solana’s AI boom and security risks

8. Januar 2026

OpenAI wants to develop audio-based AI products

OpenAI plans to introduce a new, significantly more powerful audio language model in early 2026. The goal is to enable more natural language, faster responses and smoother conversations than today's heavily text-based systems.

8. Januar 2026

AI is not a bubble, but a technological change

The existence of an AI bubble is possible. If so, however, it has been created primarily by exaggerated expectations and hype, not by the technology itself.

8. Januar 2026

The most bizarre new tech at CES 2026

The Consumer Electronics Show (CES) 2026 is showcasing not only major innovations from well-known brands, but also some particularly bizarre and surprising technologies. One particularly eye-catching exhibit is an interactive, AI-controlled panda robot that responds to touch.

8. Januar 2026

Apple’s AI strategy and development

Compared to Google, Meta and OpenAI, Apple is considered more cautious and slower when it comes to AI, as the company places greater emphasis on data protection and security. Critics see this as a disadvantage compared to its competitors.

6. Januar 2026

Costs of AI slop

Critics warn that the economic sustainability of generative AI will be put to the test in 2026. Current business models may be fundamentally flawed.

6. Januar 2026

AI coding assistants vulnerable to attack

At the 39th Chaos Communication Congress, security researcher Johann Rehberger demonstrated how vulnerable AI coding assistants such as GitHub Copilot, Claude Code and Amazon Q are to prompt injection attacks.

6. Januar 2026

Key AI contradictions

In 2026, artificial intelligence faces five major paradoxes that complicate its future impact. AI will simultaneously cause job losses and create new jobs, with human skills remaining indispensable and transitions in the labour market proceeding unevenly.

6. Januar 2026

L’Oréal uses AI for digital advertising

L'Oréal is increasingly using AI tools in its daily digital advertising production to meet the ever-growing demand for content for global markets. The technology is not intended to replace creative teams, but rather to support them.

6. Januar 2026

AI hype became more pragmatic in 2025

Last year, artificial intelligence came back down to earth. After lofty visions of imminent superintelligence, the focus shifted to concrete applications suitable for everyday use.

6. Januar 2026

2026 a turning point in the labour market

In the new year, artificial intelligence will leave its niche and bring about noticeable changes in the German labour market. Indeed's ‘Jobs & Hiring Outlook Report 2026’ shows a sharp increase in job advertisements requiring AI skills.

6. Januar 2026

AI trends in Switzerland

By 2026, artificial intelligence will have spread across various sectors in Switzerland, from healthcare to small businesses. One important milestone is Apertus, the country's first open-source language model, which is being continuously developed.

22. Dezember 2025

Switzerland grapples with AI copyright

In Switzerland, there is ongoing debate about how the use of copyright-protected content by AI should be regulated. The specific cause for concern is journalists, artists and publishers accusing AI companies of data theft.

22. Dezember 2025

What might life look like in ten years?

By 2035, AI is likely to fundamentally change everyday life and have a significant impact on areas such as health, law, work and domestic routines. In medicine, it could often serve as the first point of contact and also take over the prioritisation of treatments.

22. Dezember 2025

Europe’s AI dilemma

Europe’s challenges with artificial intelligence stem from deep regulatory, cultural, and historical differences compared with the United States. Europe produces plenty of AI talent, but many startups move to the U.S.

22. Dezember 2025

Trump’s attacks on science jeopardise ‘Genesis’

In 2025, President Trump launched the ambitious Genesis mission. This is an AI-supported project that aims to accelerate scientific breakthroughs using federal data and supercomputers, often described as a modern ‘moonshot’.

22. Dezember 2025

AI is transforming marketing workflows

Marketing agencies are increasingly integrating AI into their daily operations, not just as an add-on tool, but as a core part of briefs, production, approvals, and media optimization.

22. Dezember 2025

Skills that will shape AI careers

Economist Robert Seamans believes that future AI jobs will primarily require people who understand artificial intelligence but can also explain it in a way that is understandable to non-specialists.

18. Dezember 2025

OpenAI and ChatGPT: Rise, challenges and current milestones

Since its launch in November 2022, OpenAI's chatbot ChatGPT has evolved from a tool for writing and programming into a global AI platform with around 800 million weekly users. In 2025, the company faced increasing competition from Chinese providers.

18. Dezember 2025

AI in 2026

In the coming year, generative AI will evolve from the experimental phase to fully autonomous systems that can plan, control and execute complex workflows with minimal human supervision.

18. Dezember 2025

OpenAI’s coding agent optimises itself

OpenAI's AI coding agent Codex now independently creates most of the code it uses to improve itself. Since its introduction in May 2025, Codex has been performing tasks such as writing functions, fixing bugs and managing change proposals.

18. Dezember 2025

Opera turns browsers into workbenches

Opera, a Norwegian software company, is launching Neon, a paid, AI-powered browser. It goes beyond simple chatting and sees itself as an agent-based workspace. For $19.90 a month, it bundles leading AI models and offers three modes: Chat, Do and Make.

18. Dezember 2025

‘Always-on’ device

OpenAI and former Apple chief designer Jony Ive are reportedly working on a constantly active AI device. A prototype is already circulating internally, but it is unlikely that the ambitious vision will be fully realised in 2026.

18. Dezember 2025

Salesforce expands Agentforce

The US cloud computing company Salesforce is expanding its Agentforce 360 platform. The aim is to close the context gap for AI agents in companies. New features include deeper integration with Informatica, Data 360 and MuleSoft.

16. Dezember 2025

EU investigates Google’s use of online content

The European Union has launched a competition investigation into Google. The investigation focuses on whether the company is unfairly using online content to train its AI models, such as Gemini.

16. Dezember 2025

Steps towards a positive net energy balance for AI

As AI is driving up demand for data centres, electricity consumption could double by 2030. This would have significant consequences for power grids, costs and emissions unless growth is strategically managed.

16. Dezember 2025

AI co-pilot makes bionic hands easier to control

Researchers at the University of Utah have developed an AI co-pilot for bionic hands to make them more intuitive and easier to use. Conventional prostheses are often abandoned because controlling the individual joints and grips requires a high level of concentration.

16. Dezember 2025

Canada and EU deepen AI partnership

Canada and the European Union have decided to significantly expand their cooperation in the areas of artificial intelligence and digital services. The occasion was the first meeting of the EU-Canada Digital Partnership Council in Montreal.

16. Dezember 2025

Time Magazine names ‘AI architects’ as Person of the Year

The US magazine Time has named the ‘architects of artificial intelligence’ as Person of the Year 2025, honouring the leading minds behind the rapid rise of AI. According to the editorial team, AI shaped the year like no other development.

16. Dezember 2025

OpenAI’s rise fuels rivalries

OpenAI was founded in 2015 by Elon Musk, Sam Altman, and other tech pioneers as a nonprofit, with the goal of developing artificial intelligence for the benefit of humanity. A decade later, Musk has left the company and launched the competing startup xAI.

11. Dezember 2025

OpenAI declares “Code Red” in light of Gemini’s success

OpenAI CEO Sam Altman has declared “Code Red” for ChatGPT following the success of Google's Gemini 3. Gemini 3 gained 200 million users in three months and outperformed ChatGPT in benchmark tests.

11. Dezember 2025

EY and NVIDIA drive physical AI forward

EY is launching a structured platform in collaboration with NVIDIA to support companies in planning, testing, and deploying physical AI. This includes robots, drones, and edge devices, among other things.

11. Dezember 2025

Companies winning with AI

A study by Japanese consulting firm NTT DATA shows that only 15% of the companies surveyed are considered AI pioneers. They are characterized by clear strategies, disciplined implementation, and company-wide adoption.

11. Dezember 2025

Yoodli increases valuation with human-centered AI

Yoodli, a Seattle-based communication training startup, has reached a valuation of over $300 million following a $40 million funding round led by WestBridge Capital. Founded by former Google and Apple engineers, the platform uses AI-powered role-playing scenarios for sales, leadership, interviews, and coaching.

11. Dezember 2025

OpenAI wins purely AI-based poker tournament

OpenAI's o3 model took first place in a five-day poker tournament featuring nine leading language models, ultimately winning $36,691. The event, organized by PokerBattle.ai, tested how AI handles uncertainty, adaptation, and strategic thinking.

11. Dezember 2025

Open-source solutions as an alternative

A study by the Fraunhofer Institute shows that the German federal administration can largely rely on its own open solutions instead of large foreign platforms. Through internal developments, the state avoids critical dependencies and maintains transparency, control, and the ability to act.

9. Dezember 2025

The UK and Germany promote quantum computing

The UK and Germany are cooperating to accelerate the commercial use of quantum computing and bridge the gap between research and practical applications in computing, sensor technology, and time measurement.

9. Dezember 2025

Frankfurt’s data centers drive regional growth

Frankfurt and the Rhine-Main region have developed into a central digital hub in Europe. The GDP of the data center industry has doubled in five years, while the overall economy has grown by only 16 percent.

9. Dezember 2025

Age and geography shape attitudes toward AI

study by Cisco shows significant age and country differences in AI acceptance. Younger users are the most active in using AI, with many also completing relevant training courses.

9. Dezember 2025

A new gold rush in the American West

The former gold mining region in the Nevada desert is currently undergoing a profound transformation as technology companies build one of the world's highest concentrations of data centers there.

9. Dezember 2025

Setbacks in AI regulations affect EU and Switzerland

The EU plans to postpone and simplify parts of its AI law and the GDPR in order to give companies easier access to data and more time for implementation. Critics warn that this would weaken data protection and security.

9. Dezember 2025

AI as a demographic necessity

Fears that AI could replace jobs overlook a harsher reality: demographic change. Aging societies face a serious labor shortage that threatens economic stability.

Anthropic unveils faster, cheaper Opus 4.5

Updates from the world of artificial intelligence

Recent News

Entdecke mehr von I.AI Consulting & Training