Claude Haiku 5.5 Reaches Copilot as Local Sandboxing Goes GA

Two themes dominate the AI coding news today: a new fast, cheap model arriving everywhere at once, and agent safety moving from opt-in to default. Claude Code shipped 2.1.293 and 2.1.294 per its official changelog, GitHub made local sandboxing generally available, and Claude Haiku 5.5 is now in Copilot.

Claude Code 2.1.293 and 2.1.294

• Haiku 5.5 support (2.1.293): adds claude-haiku-5-5, now the default Haiku model on the Anthropic API, with a 1M context window and pricing of $0.10/$0.50 per million tokens ($0.50/$2.50 for prompts over 100K).
• Mods and status lines: subagentStatusLine payloads now include agentType, and mods can set isDeferred to false on $.tool.register to list a tool's schema in the prompt from the start instead of behind tool search.
• Fixes (2.1.293): a memory leak in HTTP MCP connections, lost queued messages when moving a session to the background, and a SendMessage problem when that tool was removed.
• Hook safety (2.1.294): prompt and agent hooks written as instructions, such as "Block commands that...", could previously let actions through that they should have blocked. That is fixed, and prompt hooks on Stop and SubagentStop are judged better so Claude is less likely to stop early.

If you use natural-language hooks as a guardrail, update today. That class of hook is exactly where a silent failure hurts most.

GitHub Copilot: Sandboxing GA, Haiku 5.5 and Local Models

Local sandboxing is GA. According to the GitHub announcement, it works in Copilot CLI, the GitHub Copilot app and VS Code sessions using Agent Host, on Windows, macOS and Linux, powered by Microsoft eXecution Container (MXC). You can limit which files agent commands may read or modify, control internet and local network access, restrict Git and GitHub CLI credentials, and extend the sandbox to local MCP and language servers where supported. Enterprise-managed settings can require sandboxing, and policies apply regardless of the model in use. It is included with Copilot at no extra cost.

Claude Haiku 5.5 is generally available in Copilot. Per the changelog, it targets fast, high-volume work such as subagents, quick edits and terminal tasks, and in early testing matched Claude Sonnet 5 on many coding tasks with fewer tokens and steps. It is available on Pro, Pro+, Max, Business and Enterprise across VS Code, Visual Studio, Copilot CLI, JetBrains, Xcode, Eclipse, github.com and GitHub Mobile, billed at provider list pricing. Rollout is gradual, and Business and Enterprise admins manage access through the model policy.

Copilot CLI discovers local models. From version 1.0.94-0, running /model lists models from a running Ollama instance next to your configured and cloud models. Nothing is added automatically: you review the provider and endpoint, then add it for the session or without switching. Models must support tool calling and streaming. Note that a local model does not enable offline mode, which stays opt-in via COPILOT_OFFLINE=true.

Secret detection model. GitHub also introduced a purpose-built model for leaked secrets that reads surrounding code to catch credentials without a recognizable token format, such as plain passwords. Opt-in push protection and Copilot /security-review checks consume GitHub AI Credits.

OpenAI Codex

The Codex changelog lists a ChatGPT for iOS update (October 7) that opens Codex task links directly and adds a redesigned approval picker. It also reminds teams that GPT-5.5 retires from Codex for ChatGPT sign-in on October 14, 2026 (the API is unaffected), and that gpt-6.1-sol, added September 29, offers near-Astra performance for complex work at lower cost. If you pinned GPT-5.5 in Codex, switch before the 14th.

What to do today

• Update Claude Code to 2.1.294 and re-test any instruction-style hooks.
• Turn on Copilot sandboxing for agent sessions on untrusted repositories, and have admins enforce it org-wide.
• Try Haiku 5.5 for subagent and quick-edit workloads to cut cost and latency.
• Move Codex workflows off GPT-5.5 before October 14.

Sources

• Claude Code changelog · Copilot local sandboxing GA · Haiku 5.5 in Copilot · Copilot CLI local models · Secret detection model · Codex changelog