NoCode Weekly: July 5 — Anthropic's Export Ban Ends, Claude Sonnet 5 Ships, Microsoft Spins Up $2.5B AI Company

TL;DR

U.S. lifts all export restrictions on Anthropic’s Fable and Mythos — after 18 days offline, Anthropic’s most powerful models are back in global rotation [1][2].
Claude Sonnet 5 launches — the most agentic Sonnet yet, priced at $2/$10 per million tokens, with near-Opus-level performance [3][4].
Microsoft creates $2.5B AI adoption company — new subsidiary to help enterprises deploy AI at scale [5].
OpenAI discusses government stake — the company floated a public fund seeded with AI company stocks [6].
Google Opal gains momentum — the free no-code AI mini-app builder continues expanding its user base [7].

🔥 Top No-Code & AI News

1. U.S. Fully Lifts Export Controls on Anthropic’s Fable and Mythos

The Commerce Department lifted all restrictions on Anthropic’s AI models on June 30, ending an 18-day ban that had blocked foreign nationals — including Anthropic’s own employees — from accessing the company’s most powerful frontier models [1][2].

What happened: On June 12, the government ordered Anthropic to suspend access to Fable 5 and Mythos 5 globally after discovering a jailbreak vulnerability. Anthropic complied, pulling the models offline entirely. On June 30, the Commerce Department reversed course and lifted all export controls [2][3].

Why it matters for no-code builders: Fable and Mythos power the reasoning capabilities behind AI-powered automation tools. Platforms that rely on Anthropic’s API — including n8n’s dedicated Anthropic node, Make’s AI agents, and tools like Lindy and MindStudio — will see restored access to the full model suite. If you’ve been waiting for Mythos-powered templates or workflows, the pipeline is open again [1].

2. Claude Sonnet 5 Launches — Most Agentic Sonnet Yet

Anthropic released Claude Sonnet 5 on June 30, describing it as “the most agentic Sonnet model the company has shipped” [3][4].

Key specs:

Pricing: $2 per million input tokens, $10 per million output tokens (introductory) [3][4]
Performance: Near Opus 4.8 levels on coding, reasoning, and tool use
Capabilities: Can make plans, use browsers and terminals, and run autonomously at a level that previously required Opus [4]

The tradeoff: Early reviews note that Sonnet 5’s agentic capabilities come with higher token consumption in practice. The “hidden cost” is that agentic workflows generate more intermediate steps, so actual spend can exceed the base pricing [4][8].

Why it matters: At $2/$10, Sonnet 5 is priced to compete directly with GPT-4o and Gemini Pro for agentic workflows. For no-code builders using n8n, Make, or Zapier’s AI steps, this means cheaper access to frontier-level reasoning. If you’re on Zapier’s Advanced tier (3x task multiplier for GPT-4o/Claude Sonnet), Sonnet 5’s lower base price could offset the multiplier [3].

3. Microsoft Creates $2.5B AI Adoption Company

Microsoft announced on July 2 it is creating a new subsidiary dedicated to helping enterprise customers adopt AI, backed by $2.5 billion in initial funding [5].

What happened: The new entity will provide consulting, implementation, and managed services for companies deploying Microsoft’s AI stack — including Copilot, Azure AI, and the new MAI model family [5].

Why it matters: Microsoft’s move signals that enterprise AI adoption is still a bottleneck. The company isn’t just selling tools anymore — it’s building a services arm to ensure customers actually use them. For no-code builders working with enterprise clients, this could mean more demand for Power Platform, Copilot Studio, and Azure-based automation workflows [5].

4. Microsoft Launches MAI Model Family — Reducing OpenAI Dependence

Alongside the adoption company, Microsoft unveiled its in-house MAI (Microsoft AI) model family on June 30 [9].

What happened: At Build 2026, Microsoft announced seven new AI models designed to reduce its reliance on OpenAI. The MAI models cover reasoning, coding, and multimodal tasks, positioning Microsoft as both a customer and competitor in the foundation model space [9][10].

Why it matters: Microsoft’s dual strategy — investing in OpenAI while building competing models — gives it leverage. For no-code builders, this could eventually mean more model choices within Power Platform and Copilot tools, potentially at lower cost than OpenAI-only integrations [9][10].

5. OpenAI Floats Government Stake Fund

OpenAI has discussed the possibility of a public fund seeded with AI company stocks, in which the U.S. government could take a stake [6].

What happened: The company reportedly floated the idea of a public fund that would give the government a ownership position in AI infrastructure, similar to sovereign wealth fund models. Details remain thin, but the conversation signals OpenAI’s willingness to deepen its relationship with federal policy [6].

Why it matters: If the government takes a stake in AI companies, regulatory frameworks could shift to favor domestic AI deployment. For no-code builders, this could mean fewer export restrictions (like the Anthropic ban that just ended) and more government-backed AI infrastructure — but also more scrutiny on how AI tools are used [6].

6. Google Opal: Free No-Code AI App Builder Gains Traction

Google’s Opal — a free, experimental no-code tool for building AI mini-apps — continues to expand its user base and capabilities [7].

What happened: Originally launched in July 2025, Opal lets anyone build AI-powered apps by describing what they want in natural language. The tool chains together AI models, handles hosting, and requires zero code. Google expanded access to 15 additional countries in October 2025, and adoption has accelerated through early 2026 [7][11].

Why it matters: Opal represents Google’s answer to the “vibe coding” movement. For non-developers who want to build simple AI tools — research assistants, content generators, data processors — Opal removes the need for even no-code platforms like Bubble or Glide. The catch: it’s still experimental, with limited customization and no production-grade hosting guarantees [7].

7. AMD Advancing AI 2026 — July 22-23

AMD’s annual AI infrastructure event is set for July 22-23 in San Francisco, bringing together AI developers, customers, and partners to explore the latest in AI hardware and software [12].

What to expect: AMD has been positioning its MI300X and upcoming MI400 series as cost-effective alternatives to NVIDIA’s H100/H200 GPUs. The event will likely feature new chip announcements, software ecosystem updates, and enterprise AI deployment case studies [12].

Why it matters: AMD’s growing AI hardware presence means more competition in the GPU market, which could eventually lower costs for AI inference — the compute that powers the models behind no-code AI tools [12].

⚡ Quick Hits

Anthropic’s Sonnet 5 includes cyber safeguards — the model ships with built-in protections against dangerous cyber use, a first for a mid-tier model [13].
Stanford HAI: Workers want AI for repetitive tasks, not decision-making — a new study found workers prefer to retain agency and oversight over AI systems, pushing back against full automation narratives [14].
METR: AI coding tools may not boost experienced developers as much as expected — a randomized controlled trial found early-2025 AI tools had limited productivity impact on experienced open-source developers [15].
OpenAI GPT-5.2 models retired — as of June 12, GPT-5.2 Instant, Thinking, and Pro are no longer available in ChatGPT. GPT-5.6 Sol remains limited to trusted partners [16].

📅 What’s Next

Anthropic’s global model access restoration — expect Fable and Mythos to reappear in automation platforms over the coming weeks [1].
Claude Sonnet 5 integration wave — n8n, Make, and Zapier will likely add Sonnet 5 templates and default model options [3].
Microsoft’s AI adoption company — watch for enterprise Power Platform and Copilot Studio updates tied to the new subsidiary [5].
AMD Advancing AI event (July 22-23) — potential new chip announcements that could affect AI inference costs [12].
Bubble’s Prompt AI Workflow Action — the simplified LLM step without API Connector was targeting end of June — expect it any day now [17].

Sources

NiteAgent — AI agent development, frameworks, and production patterns
ToolBrain — tool reviews, LLM comparisons, and AI workflow guides
CodeIntel Log — code quality, debugging, and software engineering benchmarks
Hermes Tutorials — Hermes Agent setup, configuration, and advanced workflows

Cross-links automatically generated from NoCode Insider.