diffnotes.tech — тех-блог о программировании, AI и DevOps

Все ai claude-code llm agents open-source productivity tips anthropic developer-tools mcp coding claude tools cursor openai gemini google api codex automation workflow ai-agents cli testing qwen ai-tools ide skills security comparison tokens voice models plugins coding-tools ai-coding cybersecurity multimodal openrouter tdd gpt coding-agent playwright pricing orchestration codex-cli multi-agent context-window coding-agents memory python stealth-models alibaba google-io-2026 moe git openclaw ralph-loop autonomous-coding ios swift xcode computer-use gpt-5.4 code-review browser-automation unity game-development xai grok context-engineering vibe-coding web-scraping chatgpt browser leak gemma hunter-alpha ai-models deepseek video-generation owl-alpha coding-assistant benchmarks benchmark research devtools deep-research terminal qa php laravel assistant worktrees docker parallel-development oauth websocket context-management mobile github copilot perplexity multi-model image-generation remotion video shorts instagram tiktok permissions future code-intelligence knowledge-graph future-of-programming opinion hooks xctest commands local-ai liquid-ai privacy fast-mode copilot-cli macos linux windows machine-learning cron scheduled-tasks effort settings godot unreal-engine search-api tavily exa agent-teams opus-4.6 expo cowork remote-control plugin google-colab responsive-design frontend telegram discord channels astral superapp kimi licensing documentation prompts figma design web-development demo gamedev gemini-cli speech scraping self-improvement ultraplan debugging function-calling free-tools china tts elevenlabs infrastructure configuration skill microsoft dotnet cost-optimization nous-research glm gpt-6 llama healer-alpha elephant-alpha gpt-5-5 tmux stealth-launch fal elixir linear rust tencent voice-cloning reasoning protocol nvidia nemotron mythos policy dense-model game-dev open-beta sonnet gpt-55 spacex managed-agents realtime subq subquadratic long-context transformers finance edge-ai vision ollama rag vector-search notion typescript workers malware chrome leaks veo lmarena fingerprinting api-pricing onboarding opus-4-8
$ ls -lt posts/ --filter=china
glm-51-open-source-swe-bench.md
GLM-5.1 — open-source модель из Китая обошла Claude Opus и GPT-5.4 на SWE-Bench Pro, кодит 8 часов без остановки и стоит в 5 раз дешевле
> · 8 мин

GLM-5.1 — open-source модель из Китая обошла Claude Opus и GPT-5.4 на SWE-Bench Pro, кодит 8 часов без остановки и стоит в 5 раз дешевле

Z.ai выпустила GLM-5.1 под MIT-лицензией: 58.4% на SWE-Bench Pro (выше Claude Opus 4.6 и GPT-5.4), 600+ итераций оптимизации без плато. API в 5-8 раз дешевле конкурентов, подключается к Claude Code за 5 минут.

ai agents llm coding
subscribe.sh

$ cat /dev/blog/updates

> Свежие заметки о программировании,

> DevOps и AI — прямо в мессенджер

./subscribe