Articles
Analysis, field reports and context on AI in processes, marketing and research.
10 articles
Claude Commerce Agents: So verbessern KI-Agenten deinen E-Commerce-Shop
Verständlicher Leitfaden zu Claude Commerce Agents für deutsche Shop-Betreiber: Was sie können, wie die Architektur funktioniert und welche Vorteile sie für Conversion und Kundenservice bieten.
Gemini 3.8 Flash und 3.8 Flash Cyber: Alle Infos zu Googles neuesten Modellen
Google hat Gemini 3.8 Flash und 3.8 Flash Cyber vorgestellt. Wir erklären die Benchmarks, Preise und was das Fairwind Program für Cybersecurity-Teams bedeutet.
KI-Angebotserstellung im Handwerk: Zeitersparnis, Workflows und realistische Erwartungen
KI-Agenten automatisieren die Mengenermittlung aus Bauplänen und sparen laut Jobber 6–8 Stunden pro Woche. Erfahren Sie, welche Tools und Hürden es gibt.
Qwen3.8-27B-OBLITERATED: Was bedeutet Abliterierung bei Large Language Models?
Abliterierung bei LLMs erklärt: Wie SVD und LEACE die Sicherheitsmechanismen von Qwen3.8-27B entfernen und was das für die AI-Safety-Forschung bedeutet.
Claude Fable 5.1: Everything About Anthropic's New Coding Model
Claude Fable 5.1 costs up to 45% less than Fable 5 on agentic tasks. Everything on benchmarks, pricing, and the new Enterprise Frontier Safeguards.
GLM-5.3-Flash unmasked: what the stealth test means for teams using coding agents
The anonymous OpenRouter model “Ox Alpha” came from Z.ai: a million-token context window, MIT license, 7.5 cents per million input tokens. What that means for running coding agents — and where the risks remain.
Using Ox Alpha in Claude Code: setup guide, limits, and the data risk
Connecting Ox Alpha to Claude Code takes just four environment variables, and a one-million-token context window is free during the preview — but prompts and responses end up with an anonymous provider. Setup guide, limits, and the GDPR risk.
What are tokens? The unit of computation behind every AI model explained
Tokens are the unit of computation behind every language model. An overview of how tokenizers break down language, why thinking costs money, and how to manage your own usage.
New model released: Claude Opus 5
Anthropic releases Claude Opus 5: nearly matching Fable 5's intelligence at half the price. The key facts on performance, cost, and safety limits at a glance.
New model released: Kimi K3 from Moonshot AI
Moonshot AI releases Kimi K3: the first open model in the 3-trillion-parameter class. The key facts on performance, architecture, and availability at a glance.
Want to put AI to work in a meaningful way?
Whether it is process automation, marketing or choosing the right tools: tell us what you are working on.
Get in touchOr email us directly: hallo@kiprozess.com