Skip to main content
Selection All articles Must reads only
Topic: AI models
3D assets 1 Agent security 5 Agent traffic 1 Agentic Commerce 2 Agentic development 1 Agentic coding 1 Agentic shopping 1 API economy 1 Architecture 4 Image generation 1 Bot detection 1 CI/CD 1 Cloud budgets 1 Cloud Native 1 Code governance 1 Code security 1 Code quality 2 Container security 1 Content quality 3 Content protection 1 Content structure 1 Databases 1 Design workflows 1 Digital sovereignty 1 Documentation 1 E-Commerce 1 Developer tooling 1 Development platforms 1 Development process 1 Frontend 1 Frontend AI 1 Frontend tooling 1 Generative Interfaces 1 IT security 1 AI & Architecture 2 AI & Teams 1 AI agents 28 AI architecture 6 AI assets 1 AI assistants 1 AI selection 1 AI automation 3 AI benchmark 1 AI operations 2 AI images 1 AI image generation 1 AI image models 3 AI coding 2 AI Commerce 1 AI compliance 1 AI adoption 1 AI development 18 AI governance 14 AI infrastructure 5 AI content 1 AI interfaces 2 AI costs 5 AI market 2 AI models 36 AI modernisation 1 AI platforms 3 AI in practice 10 AI projects 1 AI prototyping 1 AI processes 1 AI quality assurance 1 AI editing 1 AI safety 20 AI visibility 3 AI sovereignty 1 AI strategy 2 AI search 1 AI tooling 71 AI Visuals 1 AI knowledge management 1 AI workflows 3 AI collaboration 1 AI in operations 1 AI in e-commerce 1 AI in everyday development 1 AI in software design 1 AI in development 1 AI and SMEs 1 Classification 1 Legacy migration 1 Legacy modernisation 2 Local AI 1 Brand management 1 Model access 1 Observability 2 Open models 2 Open-weight AI 1 Open-weight models 2 Open Source 4 Project costs 1 Project organisation 1 Process 1 Quality assurance 4 Release processes 1 Reverse engineering 1 Security 4 Visibility 1 Software supply chain 1 Software security 1 Software architecture 1 Software development 1 Software quality 1 Search and content 1 Search visibility 1 Search strategy 1 Supply chain security 3 Team building 1 Vibe Coding 3 Voice interfaces 1 Voice AI 1 Web security 1 Web standards 1 Web strategy 1 Web security 1 Web standards 2
✕ Topic
36 articles
AI models AI
19.11.2026 · 4 min · AI

MAI Code 2: Microsoft bets on low-cost models for agents in continuous operation

Microsoft unveils seven of its own AI models. MAI Code 2 is said to achieve 98 per cent of the performance of GPT-6.1 Sol at 40 per cent of the cost – intended for AI agents that run continuously.

1 Source linked → Read more
AI models AI
19.11.2026 · 5 min · AI

Gemini 4 Neon: Google's faster model in a value-for-money comparison

Google releases Gemini 4 Neon, the fourth frontier model in seven weeks. Faster than its sister model Carbon, cheaper to run – and with a context window of one million tokens.

3 Sources linked → Read more
AI models AI
18.11.2026 · 5 min · AI

GPT-6.2 Astra and Sol: What the new OpenAI models change for projects

OpenAI has released two new models: Astra for complex project work, Sol as a cheaper everyday variant. Both are available in Codex and the API — with costs reportedly reduced by the provider.

2 Sources linked → Read more
AI models Must read AI
17.11.2026 · 5 min · AI

AI coding at a plateau: Moonshot shifts focus, Qwen4-Max lowers costs

Moonshot AI sees hardly any model progress in programming and is aligning Kimi towards the natural sciences. Alibaba cuts the cost per task by 62 percent with Qwen4-Max. The lever is shifting from benchmarks to processes.

2 Sources linked → Read more
AI models AI
13.11.2026 · 5 min · AI

Gemini 4 Hydrogen: budget model with weaknesses in agent operation

Google positions Gemini 4 Hydrogen as a fast everyday model at a lower price. Benchmarks see it close to Claude Sonnet 5.5 – however, it falls short on agent-based terminal work. What this means for development teams.

1 Source linked → Read more
AI models AI
10.11.2026 · 4 min · AI

Claude Fable 5.2 and Mythos 5.2: Speed becomes a pricing factor

Anthropic has released Claude Fable 5.2 and Claude Mythos 5.2. The models are cheaper – those who want faster outputs pay double, according to the provider. What this means for the calculation of AI-assisted development.

1 Source linked → Read more
AI models AI
10.11.2026 · 5 min · AI

Muse Ember: Meta's small model for cost-effective agent workflows

Meta positions Muse Ember as the smallest and most efficient model in its own portfolio — tailored to tool calls, routing and data extraction. According to the manufacturer, it remains weak on real programming tasks.

1 Source linked → Read more
AI models AI
10.11.2026 · 4 min · AI

Anthropic: When AI writes over 90 per cent of its own core code

With Fable 5.2 and Mythos 5.2, Anthropic is introducing a new Claude generation. According to the manufacturer, over 90 per cent of the accepted changes to its own core code already come from the model.

2 Sources linked → Read more
AI models AI
30.10.2026 · 4 min · AI

Mistral Large 4: open European model with a coding focus

Mistral has released the weights of Large 4, internally "Le Chonk". The model trained in Europe scores well in programming and in the operating cost profile, but remains behind Chinese top models in the overall comparison.

1 Source linked → Read more
AI models AI
29.10.2026 · 5 min · AI

Mistral Large 4: What an EU model should deliver for coding projects

Mistral releases Large 4, a new language model complete with weights, and explicitly positions it for coding and cyber tasks. For companies with data sovereignty requirements, this is more than a footnote.

1 Source linked → Read more
AI models AI
27.10.2026 · 5 min · AI

GPT-6.1 Terra: Which AI model suits which development task?

OpenAI halves the API prices of its predecessor with GPT-6.1 Terra and at the same time provides a rule of thumb for model selection. For projects, what will matter going forward is less the model size than the right allocation per use case.

2 Sources linked → Read more
AI models AI
23.10.2026 · 4 min · AI

Flint: Local coding model with 40 billion parameters

The start-up Recursive presents Flint, an MoE language model for local coding tasks. It was developed in four months with the help of an in-house AI system – a signal for teams with strict data protection requirements.

1 Source linked → Read more
AI models AI
20.10.2026 · 5 min · AI

Clairic: Cloudflare's lightweight model for document extraction

Cloudflare has introduced an AI model with Clairic that is intended to extract data from documents faster and more cheaply than large language models. For mid-sized companies' document processes, this is a reason to review their own model choice.

1 Source linked → Read more
AI models AI
15.10.2026 · 5 min · AI

Looped Transformers: What OpenAI's next models promise for code

OpenAI is reportedly working on a flagship model with Looped Transformers – a technique intended primarily to help with mathematics and code. One to two new models are expected to be released by the end of the year.

1 Source linked → Read more
AI models AI
14.10.2026 · 4 min · AI

Gemini 4 Argon: Google's new model for code and security analysis

Google has released Gemini 4 Argon. The model targets programming, error analysis and the autonomous detection of vulnerabilities – and costs users more.

2 Sources linked → Read more
AI models AI
13.10.2026 · 4 min · AI

Muse Wonder: Meta's new frontier models put to the agent test

Meta introduces two new flagship models with Muse Wonder and Muse Wonder Full. They land in the top three in six out of ten benchmarks — while the cost per task has nearly tripled.

2 Sources linked → Read more
AI models AI
07.10.2026 · 4 min · AI

EmbeddingGemma 2: Offline RAG with around 191 MB RAM

Google's open embedding model converts text, images, audio and code into vectors directly on the device – using around 191 MB of RAM. For search and RAG in websites and CMS, this opens up a path without cloud outflow.

1 Source linked → Read more
AI models AI
07.10.2026 · 4 min · AI

Mistral Large 4: Europe's leading model for AI-powered development

Mistral introduces Large 4, a new flagship model. For companies that want to run AI features and coding assistants independently of US providers, this noticeably expands their options.

1 Source linked → Read more
AI models AI
06.10.2026 · 5 min · AI

Reflection Beam: Open-weight model prioritises efficiency over peak scores

With Beam, Reflection releases its first open-weight model. It activates only a fraction of its parameters per token and is intended to get by with significantly less computing power for coding and reasoning.

1 Source linked → Read more
AI models AI
02.10.2026 · 4 min · AI

Three MAI audio models: Microsoft equips voice agents

Microsoft is expanding its own MAI model family with three audio models for transcription and speech output. This creates new building blocks for projects involving customer dialogue or accessibility.

1 Source linked → Read more
AI models
01.10.2026 · 4 min · AI

Gemini 4 Argon: A measured model change in the development stack

Google has introduced a new frontier model with Gemini 4 Argon. What is reliably known at launch – and the criteria by which you can decide whether a model switch makes sense in your own project.

1 Source linked → Read more
AI models AI
01.10.2026 · 5 min · AI

Gemini 4 Argon: top-level performance with higher consumption per task

Google presents Gemini 4 Argon, its first frontier model in over seven months. According to measurements, it reaches the level of GPT-6 Astra, remains behind Claude Opus 5.5 and requires significantly more computing effort per task.

2 Sources linked → Read more
AI models AI
30.09.2026 · 5 min · AI

Sonnet 5.5 and GPT-6.1 Sol: Mid-range becomes the standard tool

Anthropic and OpenAI are unveiling models almost simultaneously that come close to their flagship models but are intended to operate more cheaply. For teams with AI coding assistants, this shifts the question of model selection.

2 Sources linked → Read more
AI models AI
29.09.2026 · 4 min · AI

Claude Sonnet 5.5: Leap in coding benchmarks at lower cost

Anthropic expands the Claude 5.5 family with Sonnet 5.5. In the Terminal-Bench coding benchmark, the score rises from 10.3 to 70.6 per cent – at lower cost than the predecessor.

1 Source linked → Read more
AI models AI
29.09.2026 · 4 min · AI

Claude Sonnet 5.5: What a faster mid-range model delivers in everyday project work

Anthropic has introduced Claude Sonnet 5.5, the second model in its current 5.5 family. It is said to work faster and more cheaply than its predecessor – for coding assistants and automations, that is precisely the decisive factor.

1 Source linked → Read more
AI models AI
23.09.2026 · 5 min · AI

GPT-6 and Claude Opus 5.5: more performance at significantly lower prices

OpenAI announces half as many errors with GPT-6 at halved API prices, Anthropic introduces Claude Opus 5.5 around 40 per cent cheaper. For AI features in websites and internal tools, this shifts the calculation.

2 Sources linked → Read more
AI models AI
22.09.2026 · 5 min · AI

Grok 4.7 compared: How much model quality does agentic coding need?

xAI has unveiled its strongest model to date with Grok 4.7. In the Artificial Analysis Intelligence Index it sits mid-table with 46 points, and the gap is wider in agentic coding. On the other hand, it is inexpensive — a trade-off for development teams.

1 Source linked → Read more
AI models Must read AI
11.09.2026 · 5 min · AI

Deepseek V4.1-Flash lowers the cost of AI coding agents

Deepseek releases V4.1-Flash under the MIT licence: a quarter of the memory requirement in the KV cache and coding results at a top level. Agent workflows in development projects thus become computationally attainable.

1 Source linked → Read more
AI models AI
04.09.2026 · 5 min · AI

GPT-6 Astra: Why OpenAI gives full access to only a few companies

OpenAI introduces GPT-6 Astra, a model that the company itself classifies as critical in the field of cybersecurity. Initially, only selected companies are to receive full access, while a pared-down version will apparently become publicly available.

2 Sources linked → Read more
AI models AI
04.09.2026 · 5 min · AI

Agentic AI models: What Muse Spark 1.3 means for project budgets

Meta has released Muse Spark 1.3, the fourth model in its series within five months, undercutting the competition on price per task. Claude Fable 5.1 remains ahead on quality. For projects with AI agents, the choice of model thus becomes an architectural question.

2 Sources linked → Read more
AI models AI
04.09.2026 · 5 min · AI

GPT-6 Astra: What OpenAI's new model means for software projects

OpenAI introduces GPT-6 Astra: faster and cheaper on complex coding tasks, while being classified as critical for the first time. What this means for development teams and their review processes.

2 Sources linked → Read more
AI models AI
03.09.2026 · 5 min · AI

Gemini 3.8 Flash: stronger coding, more expensive tasks

Google releases Gemini 3.8 Flash, the third Flash model in six weeks. It makes significant gains in agentic coding – but uses around 30 per cent more output tokens per task.

1 Source linked → Read more
AI models
02.09.2026 · 5 min · AI

Claude Fable 5.1 and Mythos 5.1: more agentic coding, lower costs

Anthropic introduces two new models: significantly stronger at agentic coding, up to 45 per cent cheaper to run and with relaxed safety filters. What this means for the economic viability of AI in development projects.

2 Sources linked → Read more
AI models AI
27.08.2026 · 5 min · AI

Qwen3.8-Flash-Next: Price pressure on AI coding assistants

Alibaba introduces Qwen3.8-Flash-Next, a mixture-of-experts model that, according to benchmarks, outperforms larger competitors on coding tasks — at a fraction of the training cost. For project budgets, that changes the calculation.

1 Source linked → Read more
AI models AI
18.08.2026 · 5 min · AI

Open models, large models, small models: sorting out the situation

Chinese language models are growing, American ones are opening up, and new candidates are emerging from Switzerland and Germany. For web and tool projects, this shift will determine where AI functions run in the future.

1 Source linked → Read more
AI models AI
13.08.2026 · 5 min · AI

Weekly model changes: what this means for code workflows

Gemini 3.7 Flash, Deepseek V4-Pro and Grok 4.6 are launching within a few days of each other and are shifting the price-performance ratio for code and agents. At the same time, market data shows that expensive top-tier models are finding few buyers.

5 Sources linked → Read more

Barrierefreiheit

Passen Sie die Darstellung an Ihre Bedürfnisse an. Ihre Auswahl bleibt auf diesem Gerät gespeichert.

Schrift und Text

100 %

Schriftgröße der Seite

Farben und Kontrast

Lesen und Bedienen