AI News, 21 May 2026

By SmartBolig.net Editorial · 21 May 2026 · 6 selected sources

Today’s roundup is led by Google’s new Gemini 3.5 model series, unveiled at I/O, while GPT-5.5 has, per a brief OpenAI headline, surfaced inside Databricks’ enterprise agent tooling. Four terminal-based AI tools - OpenClaw, Codex, Claude Code and Gemini CLI - also picked up new version numbers, though none of their changelogs made it into the material gathered here.

Editorial method

Candidates are checked against recent URL, topic, and source-set duplicates. The original sources remain authoritative.

Today’s Stories

1. OpenAI: Databricks brings GPT-5.5 to enterprise agent workflows

What changed: OpenAI published the update on 14 May 2026. Per the headline from OpenAI, GPT-5.5 is now wired into Databricks tooling for enterprise agent workflows. The source gives no description or examples of how the integration works.

Why it matters: The direct relevance for home users is small, but it signals GPT-5.5 continuing to spread beyond OpenAI’s own products into third-party platforms that companies build internal AI agents on.

How to verify it: Check Databricks’ own model catalog to confirm whether and how GPT-5.5 is actually available in your instance.

Uncertainty: There’s nothing on pricing, which Databricks tier is covered, regional availability, or what ‘agent workflows’ concretely means.

Source: Databricks brings GPT-5.5 to enterprise agent workflows

2. Google AI: Gemini 3.5: frontier intelligence with action

What changed: Google AI published the update on 18 May 2026. Google introduced Gemini 3.5 as its newest model series at I/O, describing it as pairing advanced intelligence with the ability to take action. The rest of the page is navigation menus and adds no further concrete detail.

Short source excerpt: At Google I/O we released Gemini 3.5, our latest series of models combining frontier intelligence with action. Gemini 3.5: frontier…

Why it matters: If you use the Gemini app or build on the Gemini API, this points to a new model generation that Google itself frames around acting on its own initiative - relevant for anything from automated routines to smart-home integrations.

How to verify it: Check the model picker in the Gemini app or Google’s model documentation to see whether 3.5 has rolled out to your account.

Uncertainty: The source shows no benchmarks, pricing, rollout timing, or what ‘action’ concretely means beyond Google’s own phrasing.

Source: Gemini 3.5: frontier intelligence with action

3. OpenClaw: openclaw 2026.5.18

What changed: OpenClaw published the update on 19 May 2026. OpenClaw pushed out version 2026.5.18 of its personal AI assistant. The only thing the source confirms is the version number and the project’s tagline about working across operating systems.

Short source excerpt: Your own personal AI assistant. Any OS. Any Platform. The lobster way. 🦞 - Release openclaw 2026.5.18 · openclaw/openclaw Release…

Why it matters: OpenClaw users who self-host the assistant get an update to pull - but nothing in the material says what it actually changes.

How to verify it: Open the actual release page on GitHub and read the changelog text before updating a system that’s in active use.

Uncertainty: The changelog text is entirely missing, so bug fixes, new features, and any breaking changes to existing setups are unknown from here.

Source: openclaw 2026.5.18

4. OpenAI Codex: 0.132.0

What changed: OpenAI Codex published the update on 19 May 2026. Codex version 0.132.0 is out from OpenAI. The source page confirms only the version number; the actual change text is not part of the material collected here.

Short source excerpt: Lightweight coding agent that runs in your terminal - Release 0.132.0 · openai/codex Release 0.132.0 · openai/codex · GitHub /…

Why it matters: Terminal Codex users get a new build to install, but whether it changes agent behavior, default models, or fixes bugs can’t be seen here.

How to verify it: Read the release notes directly on Codex’s GitHub page before updating a workflow you depend on.

Uncertainty: It’s not stated whether this is a minor fix or an update with behavior changes, nor whether it affects compatibility with existing scripts.

Source: 0.132.0

5. Claude Code: v2.1.141

What changed: Claude Code published the update on 12 May 2026. Claude Code has been bumped to v2.1.141 by Anthropic. The material gathered here contains only the version tag, not the actual description of the changes.

Short source excerpt: Claude Code is an agentic coding tool that lives in your terminal, understands your codebase, and helps you code faster…

Why it matters: Existing Claude Code users will see an update available, but what it means for daily terminal work isn’t stated in the material available.

How to verify it: Go to GitHub and review the actual changelog before assuming the update is routine.

Uncertainty: Whether this is a plain bug fix or an update with behavior changes relevant to daily use can’t be determined from the material.

Source: v2.1.141

6. Gemini CLI: Release v0.42.0

What changed: Gemini CLI published the update on 11 May 2026. Another version bump has landed: Gemini CLI is now at v0.42.0. The source confirms only the version number for this open-source terminal tool, not what the update contains.

Short source excerpt: An open-source AI agent that brings the power of Gemini directly into your terminal. - Release Release v0.42.0 · google-gemini/gemini-cli…

Why it matters: The same goes for Gemini CLI: a new version is available, but the material says nothing about whether it fixes bugs, adds features, or changes default behavior.

How to verify it: Go through the changelog on Gemini CLI’s GitHub page directly before updating in an environment that’s in active use.

Uncertainty: The scope of the changes, and whether anything behaves differently than in 0.41.x, isn’t shown in the material collected here.

Source: Release v0.42.0

Why It Matters

• The focus is changes that can affect AI products, models, browser experiences, agents, API usage, privacy, pricing, or security.
• A new feature is useful in the home only after access, logging, and rollback have been tested.
• The article separates documented information from editorial judgment and open questions.

Sources and Further Reading

Publisher	Source	Date
OpenAI	Databricks brings GPT-5.5 to enterprise agent workflows	14 May 2026
Google AI	Gemini 3.5: frontier intelligence with action	18 May 2026
OpenClaw	openclaw 2026.5.18	19 May 2026
OpenAI Codex	0.132.0	19 May 2026
Claude Code	v2.1.141	12 May 2026
Gemini CLI	Release v0.42.0	11 May 2026

Editorial Note

SmartBolig.net used automated research, deduplication, and drafting for this issue. Editorial rules require a primary source, verifiable links, and explicit uncertainty.