01The model radar
Details on the two latest updates. Open a radar item for its announcement.
Google / 13 OctDeveloper rollout
Gemini 4 Argon
A model built for long coding tasks, including migrations across a codebase.
What changed
What’s new
Google opened developer access on the paid Gemini API and in Antigravity for Google AI Ultra subscribers.
Why it matters
Try it on a multi-file migration with tests that can verify the result. The relevant question is whether it completes your repository task reliably.
Access Paid Gemini API; Antigravity for Google AI Ultra subscribers.
Primary source ↗Meta / 12 OctNew model family
Muse Wonder + Wonder Full
Models for coding and personal tasks that recall context and work through blockers.
What changed
What’s new
Meta launched two related models emphasizing retrieval of relevant memories and sustained execution across difficult tasks. The family name groups Wonder and Wonder Full.
Why it matters
Evaluate continuity across sessions and recovery from a failed approach, alongside code correctness.
Access Wonder: Meta Model API and Muse Code. Wonder Full: Muse Code subscribers.
Primary source ↗02Open source. Ready to explore.
Repositories with a reason to watch or try them.
The software is open source. Model access and hosted services can carry separate costs.
Pre-release updateWrite repository tasks in Markdown and run an AI agent through GitHub Actions.
What changed + how to try
What’s new
This pre-release tightens MCP controls, rejects unsafe pull_request_target roles in strict mode and changes model-routing rules.
Why it matters
Existing workflows may need changes before recompilation; review the release before upgrading.
Try it
GitHub CLI, a repo you can edit, Actions enabled and supported model authentication.
Agent Canvas coordinates coding agents and automations across execution backends.
What changed + how to try
What’s new
v1.27.0 adds OpenCode and Pi as local ACP agents; v1.27.1 fixes Docker bridge and theme issues.
Why it matters
A useful watch if you want one control center for several coding agents.
Try it
Self-host locally or with Docker; bring the model or agent access you want to use.
An open-source terminal coding agent with a beta desktop app and configurable model providers.
What changed + how to try
What’s new
Adds explicit upgrade paths toward OpenCode 2, an /update command and tighter terminal-session checks. This version is not the OpenCode 2 release.
Why it matters
Track it for provider choice, permission changes and how the major-version migration develops.
Try it
Install the CLI or desktop app, then connect your model provider.
A stateful agent harness that can delegate coding work to an OpenCode worker.
What changed + how to try
What’s new
Adds explicit OpenCode workers and a cloud-only webhook-management skill. The repository already existed; the worker integration is the fresh capability.
Why it matters
Try persistent coordination between harnesses. Workers must be invoked explicitly and OpenCode must be installed on PATH.
Try it
Local mode is available with your provider. Webhook management requires Letta Cloud.
04Research watch
Recent papers and publications, with their findings and limits.
13 Oct 2026 · v1arXiv preprint
Same Attack, Different Judge
Tests how Jev, LLM judges and safety classifiers handle the same adversarial inputs across moderation, routing and phishing tasks.
Why it matters
Test an agent’s routing or safety judge on your own workload, including benign inputs and false alarms.
09 Oct 2026 · v1arXiv preprint
AgentSWE
Benchmarks whether coding agents can create, edit and improve other agents that meet real workflow requirements.
Why it matters
Freeze the delivered agent and test unseen acceptance cases, including its runtime behavior.
TODAY’S EXPERIMENT · 30 minutes after setupTry github/gh-aw: triage three sample issues
GitHub Agentic Workflows runs repository tasks from Markdown instructions through GitHub Actions. Give it a crash report, confusing setup instructions and a feature request; it should produce one linked triage report.
Why it helps See whether an agent can save time sorting incoming issues and deciding what to investigate first. Compare its classifications, priorities and follow-up questions with your own judgement.
Try the step-by-step lab →The custom workflow compiled successfully with official v0.89.21 and v0.91.12 binaries. An actual GitHub Actions/model run has not been performed.EDITOR’S NOTESelected for material changes and practical relevance. New model, new capability and existing-product update are labelled separately. Recommendations are editorial; availability follows the linked primary sources.