AI Engineering Blog | Felix Schmidt

Surviving Session Crashes: How Anthropic, OpenAI, xAI and Google Handle Agent State in 2026

What happens to your carefully built context when the machine reboots or your agent session crashes? A practical comparison of the native mechanisms — Memory Tool, Responses API, store_messages, Memory Bank — plus the hybrid pattern that holds up best in practice.

AI Models

GPT‑5.5 vs. Claude Opus 4.7 vs. China: Who Leads the Field in April 2026?

In three weeks, Anthropic, OpenAI, Moonshot, DeepSeek and Z.ai shipped five top‑tier models. A sober comparison of strengths, weaknesses and costs across GPT‑5.5, Claude Opus 4.7, Kimi K2.6, GLM‑5.1 and DeepSeek V4‑Pro — with quotes from primary sources and a benchmark chart.

AI
Claude Managed Agents: Run Autonomous AI Agents Without the Infrastructure Headache
Apr 12, 2026
Artikel lesen →
AI
Time & Material Is Almost Dead. Here's What Replaces It.
Apr 9, 2026
Artikel lesen →
Developer Tools
Google Stitch + Claude Code via MCP: From Design to Code Without Copy-Paste
Mar 31, 2026
Artikel lesen →

Diese Website verwendet Drittdienste (Google reCAPTCHA, Calendly), die Cookies setzen können. Mehr dazu in meiner Datenschutzerklärung .

AI, Code & Architektur — aus der Praxis.

Anthropic Rents Colossus 1 – The Surprising Compute Deal with SpaceX/xAI

Surviving Session Crashes: How Anthropic, OpenAI, xAI and Google Handle Agent State in 2026

GPT‑5.5 vs. Claude Opus 4.7 vs. China: Who Leads the Field in April 2026?

Claude Managed Agents: Run Autonomous AI Agents Without the Infrastructure Headache

Time & Material Is Almost Dead. Here's What Replaces It.

Google Stitch + Claude Code via MCP: From Design to Code Without Copy-Paste