New agent tools ranked by what they're actually good for, with a short verdict on each. Rankings are earned, never bought.
Prompt-to-app builders
5Agents that edit real codebases
8Frameworks & SDKs
–Workflows without code
–Agents that find and synthesise
4Generation & editing
–Visuals and UI
–Speech, music, sound
–Agents that click for you
–Content and copy
1Customer-facing agents
–Agents for numbers
Releases and updates from the last two weeks. Verdicts are provisional until we've tested the tool ourselves.
Verdict: A useful debugging release for teams running crews in production: the new eval command targets the hardest question, where a run went wrong. Mostly fixes otherwise.
Test notes: Not tested: the verdict is based on the linked public release notes only.
Pricing: Open source
Verdict: A strong pick for TypeScript teams that need agents to run code in sandboxes. The new Factory pieces matter most if you run many sandboxed sessions.
Test notes: Not tested: the verdict is based on the linked public release notes only.
Pricing: Open source
Verdict: A careful, approval-first coding agent and an easy pick for VS Code users who want to try another model. It's early-stage: desktop apps and a CLI haven't shipped yet.
Test notes: Not tested: it needs an account, a paid plan or a heavy local setup. The verdict is based on the linked public sources only.
Pricing: Free extension; usage billed via API key
Verdict: A maintenance release: mostly model-catalog and default changes. Check your model selection after updating, since defaults changed for many providers.
Test notes: Not tested: the verdict is based on the linked public release notes only.
Pricing: Open source (you pay for model usage)
Verdict: Worth a look if you're building multi-agent workflows for security operations. Its README positions it as a SecOps platform, not a general-purpose framework. We checked the repo and release notes but didn't install it (it needs uv, Node 22, and agent-browser).
Test notes: Not tested: it needs an account, a paid plan or a heavy local setup. The verdict is based on the linked public sources only.
Pricing: Open source
Verdict: It's the new model to try for long refactors and migrations if you already pay for Copilot or AI Ultra. Access on lower Copilot tiers is limited.
Test notes: Not tested: it needs an account, a paid plan or a heavy local setup. The verdict is based on the linked public sources only.
Pricing: Paid (Gemini API paid tier, AI Ultra, Copilot Pro+ and up)
Verdict: A solid release if you build on Google's stack: human-in-the-loop approvals and tool-output filtering are practical safety controls. The 40% speed-up is Google's own claim; we haven't measured it.
Test notes: Not tested: the verdict is based on the linked public release notes only.
Pricing: Open source
Verdict: A good pick for short social clips with sound built in. The duration cap makes it a clip tool, not a storytelling tool.
Test notes: Not tested: it needs an account, a paid plan or a heavy local setup. The verdict is based on the linked public sources only.
Pricing: 50% off on Civitai through October (base price not stated in source)
Verdict: The cleanest typed-Python agent framework we've tried. In our offline test, tool calls and structured output worked on the first run with no API key, using the built-in TestModel. Live-model quality isn't tested yet.
Test notes: Hand-tested on 15 Oct 2026: installed pydantic-ai-slim==2.56.0 into a clean Python 3.13 venv, built an agent with one tool and a typed Pydantic output, and ran it with the built-in offline TestModel (no API key). The tool was called and a validated structured output came back in about 0.01 s. Downside: it prints a large observability/Logfire banner on run unless PYDANTIC_AI_NO_BANNER=1. Not yet run against a live LLM, so output quality is scored conservatively.
Pricing: Open source
Verdict: It's for e-commerce brands, not solo builders.
Test notes: Not tested: it needs an account, a paid plan or a heavy local setup. The verdict is based on the linked public sources only.
Pricing: Demo-led (pricing not stated)
Verdict: Interesting if you want to wrap the Claude and Codex agent SDKs in a durable, multi-replica service. Niche for simple single-process agents.
Test notes: Not tested: the verdict is based on the linked public release notes only.
Pricing: Open source
Verdict: A strong new contender for heavy, multi-file agent work. The best version sits behind the Power and Maximum Usage plans.
Test notes: Not tested: it needs an account, a paid plan or a heavy local setup. The verdict is based on the linked public sources only.
Pricing: API $4/$20 per 1M tokens; Full via subscription
Verdict: The clearest "agent as the editor" take on business video so far. It's aimed at explainers and marketing, not cinematic work.
Test notes: Not tested: it needs an account, a paid plan or a heavy local setup. The verdict is based on the linked public sources only.
Pricing: Credit-based; launch credits for every customer
Verdict: Worktree support is the headline: it makes running parallel agent tasks on one repo cleaner. Worth updating if you run Codex on several tasks at once.
Test notes: Not tested: the verdict is based on the linked public release notes only.
Pricing: Open source (needs a ChatGPT plan or API key)
Verdict: A small but useful release for anyone building skill-based subagent setups on LangGraph. In our hands-on smoke test it installed cleanly and ran out of the box, and the new skill pinning is in the code. Live-model quality isn't tested yet.
Test notes: Hand-tested on 14 Oct 2026: installed deepagents==0.7.23 from PyPI into a clean Python 3.13 venv, built an agent with create_deep_agent(), and ran it end to end against an offline stand-in model (no API key). It installed and ran cleanly, and pinned_skills is present in the skills/subagent middleware. Not yet run against a live LLM on our standard task, so output quality is scored conservatively.
Pricing: Open source
Verdict: High-end quality at a low per-second price for creators who need reference consistency. It's still a preview.
Test notes: Not tested: it needs an account, a paid plan or a heavy local setup. The verdict is based on the linked public sources only.
Pricing: From $0.014 per second (launch pricing)
Verdict: Interesting for agent pipelines thanks to the CLI. Its planned HiDream-O1-Video model integration hasn't shipped yet.
Test notes: Not tested: it needs an account, a paid plan or a heavy local setup. The verdict is based on the linked public sources only.
Pricing: Not stated in source
Verdict: A patch release with no new features. Listed for completeness; update when convenient.
Test notes: Not tested: the verdict is based on the linked public release notes only.
Pricing: Open source
The week's new agent tools with a verdict on each, every week. Free. We'll email you a link to confirm, and you can unsubscribe any time. See our Privacy Policy.
We use only essential storage by default. With your OK, we'd also use privacy-friendly analytics to see which pages help. Cookie Policy