ONDŘEJ PECH /AI
PortfolioServicesReferencesBlogContact

Blog

September 2026
  • September 22
    2026
    September 22 2026

    Grok 4.7: updating long-running agents after Grok 4.6

    ↻ Updated

    Grok 4.7 builds on 4.6. Migration needs attention to the 200k-token pricing threshold, caching and preservation of encrypted reasoning history.

    Realistické noční pracovní místo s obrazovkami zobrazujícími dlouhý agentní workflow
    Read more→
August 2026
  • August 27
    2026
    August 27 2026

    GLM-5.3-Flash: open weights, 1M context, and a cheaper agent model

    ↻ Updated

    GLM-5.3-Flash combines public MIT weights, native vision, a 1M-token context window, and much lower API pricing. For companies, it is a credible primary-worker candidate for verifiable coding, document, and multimodal workflows.

    Pracovní stůl vývojářského týmu s kódem, terminálem a lokální serverovou jednotkou pro AI coding agenty
    Read more→

Articles: 2

Models & releases ×coding agenti ×Clear filters

Categories

All articles8Models & releases2Agents & automation7AI development5Practical examples1Image, video & voice1

Tags

AI15AI agenti11automatizace7Claude7LLM6Anthropic5API4coding4model router4model routing4Gemini3Google3OpenAI3coding agenti2n8n2open-source2Opus 5.52prompt caching2RAG2reasoning2agentni workflow1AI video1Aleph Alpha1Alibaba1API migration1API pricing1Automatizace1benchmarky1context compaction1DeepSeek1enterprise AI1Fable 5.11Gemma1Genie 41GLM-5.21Google AI1Google DeepMind1GPT1GPT-5.41GPT-6 Astra1GPT-6.1 Terra1Grok 4.61IBM1Kimi K31Kolibri1lokální AI1long-context1Meta1MiniMax1Moonshot AI1multimodal1multimodalita1novinky1NVIDIA1open model1open models1Opus 51Qwen1Sonnet 5.51V4 Pro1xAI1
© 2010 – 2026 Ondřej Pech
ondrej-pech.cz·Photographer·Web Development·Agentic Engineering
InstagramLinkedInGitHub